Hugging Face is an AI infrastructure company. The Microduck is a serious robotics platform: fifteen motors, a camera, lidar, reinforcement learning. And yet the reason it’s selling like this might have very little to do with the spec sheet.

I’ve said it before: there’s a weird crossover happening between characters and robots, and the companions with real character are the ones winning. The Microduck is the cleanest case yet, because the playbook behind it is a character playbook, beat for beat. Make a distinctive character. Make entertaining content around it. Give it real personality. Then open the sandbox.

Let’s waddle through it.

Beat one: a distinctive character with personality

The duck wasn’t a branding decision bolted on at the end. Pollen Robotics (the French team Hugging Face acquired in 2025) says the duck emerged from the robot itself: the proportions, the beak, the waddling gait. They kept it because it “makes the robot approachable without hiding the mechanics underneath.”

And the family tree is even better. Disney built the BDX droids, the little waddling characters that roam their parks, and roboticists fell in love. Pollen engineer Antoine Pirrone built Open Duck Mini, a DIY, 3D-printable homage to the BDX droid that became a quiet sensation in open robotics circles.

Microduck is that lineage productized. The best character robots in the world came out of Disney, the entertainment company. The open-source world spent two years reverse-engineering the feeling, and now you can buy it for $399.

Marcel says: a duck descended from Disney royalty. I descended from a prompt. We try to avoid talking about it.

Beat two: entertaining content, not demos

Robotics companies make demo videos. Sterile labs, plain backdrops, a robot doing a task (you know, this stuff) . The Microduck launch content is entertainment: the duck waddles, roller skates, falls over, picks itself back up, and keeps going.

That last one matters. A robot that falls down is a failed demo. A character that falls down and gets back up is comedy, and it’s charm, and it’s the entire point. Pollen designed for it: the duck is 25 cm and under 800 grams, so a failed experiment is “a little robot on the floor, not a major incident.” They built a robot whose failures are watchable.

Marcel says: I know what I want for Christmas!

Subscribe now

Beat 3: open the sandbox

This is where it goes past what most brands would ever dare. The whole stack is open source: the SDK, the simulation environment, the reinforcement learning training code, all on GitHub. Pollen says the interesting list of behaviors “will come from the community.” They want new walking styles and tricks to be shared and remixed the way AI models are shared on Hugging Face.

And the community showed up immediately. People are 3D-printing their own parts and accessories. People are hacking the sound system to give their ducks new voices. People are building entire little worlds and sets for their ducks to live in.

Most brands stop at inviting fan art. This is inviting fan behavior, fan bodies, fan voices. The character isn’t just something you watch, it’s something you get to author. That’s the same instinct behind every great character universe, taken to its logical end: the fans don’t just love the world, they build it.

Marcel says: Stef didn’t even complete my training

So what

A robotics company just ran a better character launch than most entertainment brands will this year, and the market answered in hours. The lesson isn’t “make a duck.” It’s that character and product are merging. The distinctive shape got the attention. The entertaining content spread it. The personality made people care. The sandbox turned buyers into builders.

If you make characters, your new competition includes robotics companies. If you make robots, or apps, or anything that wants a place in someone’s daily life, the companions with real character are winning. Plan for personality the way you plan for features, because the Microduck just proved personality ships.

Marcel says: a duck with fifteen motors made a million dollars in six hours. I have zero motors. Stef. We need to talk.

Want your own brand character?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - the only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

Two things are happening at once. New characters are being born, and old ones are revived. And every new mascot arrives with a launch. Sometimes that just means distributing the ad. Sometimes the brand builds a whole little world around the arrival. Crocs and Dolmio just showed how good that can look.

Crocs hatched a crocodile in public

The crocodile has been on the side of the shoe since 2002. Twenty-three years as a silhouette, zero as a character. On August 13, Crocs finally introduced him: Niles, six feet tall, silent, and built like a sitcom character. Their words, not mine. Personality, interests, flaws, and ambitions.

But look at how they did it. For a month before the reveal, a three-foot egg kept showing up in weird places: TikTok Shop livestreams, platform logos, brand assets. Then a 24-hour countdown. Then a live hatching on their socials. Now a weekly micro-series filmed at Crocs HQ with real employees, where Niles hatches early and frantically tries to reassemble his own egg before the countdown expires.

@crocseveryone welcome our newest employee @Niles! you've seen him on your favorite clogs, and now he's coming to a phone screen near you (and maybe a swamp?) be sure to follow Niles as he navigates his journey as a working professional crocodile @nilesthecroc

Tiktok failed to load.

Enable 3rd party cookies or use another browser

The launch wasn’t marketing for the character. The launch was the first episode. Crocs turns 25 next year, and they’re clearly building Niles as long-term IP with full on episodes, not as a campaign asset they’ll retire in Q2.

Subscribe now

Dolmio brought the family home

Dolmio’s puppet family sold pasta sauce for decades until the brand cut the strings in the mid-2010s and went for “real-life situations.” Regular humans, regular dinners. You’ve never heard anyone quote those ads.

Now, for the brand’s 40th anniversary, the puppets are back. Agency T&P, a budget in the millions, Dolmio’s most ambitious campaign in over ten years, built on the old platform: when’s your Dolmio day. Hero film, out-of-home, a partnership with author Giovanna Fletcher, influencer content.

Dolmio was smart enough to lean on the glory days of the characters for the re-launch. They teased the family on Instagram, and the comment section did the remembering for them. Forty years of banked fame, and the campaign spends it instead of pretending it doesn’t exist.

Marcel says: they kept an entire family in storage for eight years. I have reread my contract twice since learning this.

Stanley gave a toolbox a mouth

On August 3, Stanley launched “All build. No bull.” with Stan: a talking toolbox with jobsite mileage, starring in ten spots where he trashes overengineered gizmos, empty marketing claims, and general category nonsense. The audience is construction pros, not TikTok teens.

The clever bit: a brand saying “most tool marketing is bull” sounds defensive. A toolbox saying it is funny. The character gives Stanley permission to attack its own category without sounding bitter. And it’s proof the mascot boom isn’t just a DTC cuteness play. This is a 180-year-old tool brand deciding a character is the sharpest way to talk to professionals.

Marcel says: Stan holds tools for a living and complains the whole time. We are getting lunch.

The Wall Street Journal noticed

When the WSJ writes about your niche, the niche is over, or it’s just getting started. Their piece counts the new class: Microsoft’s Mico for Copilot, Apple’s little Finder guy, Mozilla turning the Firefox fox into a full character named Kit, Reddit’s refreshed Snoo. Even Penguin’s bird is getting more screen time.

The pattern: the more abstract and mistrusted the product, the cuddlier the messenger. AI assistants, browsers, water filters. Lancaster University’s Anthony Patterson puts it plainly: mascots give “a voice, a personality, a face even” to companies that feel cold and impersonal. The piece also cites research that brands with mascots were 37% more likely to grow market share, and notes Duo the owl now has over 20 million followers. A mascot is a reputation you actually control. Celebrities get scandals. Characters get merchandising.

Marcel says: The Journal calls us “cuddly.” I prefer “hot”

Spotted in the wild

Two more from the pile. Zespri is running “The Zing” in Belgium, a healthy-habits platform fronted by the Kiwi Brothers, the fuzzy kiwifruit characters that became a phenomenon in Japan and are now quietly going global as the brand’s rebels-for-healthy-eating.

And SpringWell, a water filtration brand, answered the question “if water could talk, what would it say?” with: it would sing a ballad. Quality Meats Creative made it in stop-motion, contaminants and all.

When tap water gets a singing voice, the era is officially on.

So what?

If you’re bringing a character into the world, there’s a spectrum. The floor is having a mascot. In between is making a fun ad with it. The ceiling is what Crocs and Dolmio did: build a small world around the arrival. Give a new character a birth, like Niles. Give a returning one a homecoming, like the Dolmio family. And plan the second episode before the first one airs, because a character with no next appearance is just a logo with eyes.

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - the only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

At some point you have to eat your own cooking.

So: Marcel has a show now.

Introducing Character News Network

It’s a news desk for character news. Marcel anchors. I’m the sidekick. Every episode covers what this newsletter covers: mascots, character IP, world building, the technology moving it all, except now a blue dog with a skull charm delivers it to camera, and I get talked over at my own desk.

Episode one: the bear that stood still for 35 years

The Ralph Lauren Polo Bear has been on knitwear since 1991. Embroidered, illustrated, instantly recognisable, and completely static. Thirty-five years of standing very still on very expensive sweaters. The definition of a distinctive asset used as decoration.

That just changed. Ralph Lauren has started building an actual world around the bear. It has scenes now. A personality. Situations, things that happen to it. There’s an animated short.

The bear went from decoration to IP.

What the movie star is actually for

The celebrity is not the point. The celebrity is the on-ramp.

You can’t own a movie star. You can’t control one either, and as Marcel cheerfully pointed out on the show, they don’t last forever. The Michelin Man is 128. Ronald hasn’t aged since the seventies. Characters compound; celebrities depreciate.

But a celebrity is very good at one thing: attention, right now. So Ralph Lauren rented that attention once, and spent all of it walking people into the bear’s world. The thing they own forever. Most brands rent the attention, run the ad, and then rent a different celebrity next year. Nothing compounds. Ralph Lauren used the rental to build equity in an asset that’s been on the balance sheet since 1991.

Marcel says: the man is a doorman for the bear universe. A well-dressed doorman.

Behind the desk: Marcel is live

One thing you should know about the show: Marcel is not an AI-generated video in this show. He’s a live puppet.

I worked with a 3D artist who rigged him as a VTuber. A full 3D Marcel with a working face, driven by a face recognition setup I vibecoded: a webcam watches me, the software reads my expressions, and Marcel performs them in real time. I raise an eyebrow, he raises an eyebrow. I talk, he talks. I look unimpressed at my co-host; well, you’ve seen the episode (hopefully).

Here’s why that mattered for me. Every other way of animating Marcel puts something between the performance and him. A render queue. An animator. AI hallucinations. An editor. Hours of waiting. Real-time puppeting removes all of it. The performance goes into the character the moment it happens. The distance between “I have a take” and “Marcel has a take” is zero.

Live puppeting is the most direct way there is to give a character life. It’s also why this show can exist without a studio budget: one person, one webcam, one blue dog who never misses a cue.

Marcel says: I’m told a man is performing me. I’ve never seen him.

Why this needed to be a show

Because the whole thesis of this newsletter is that a logo sits still while a character does things, and a character gets more valuable every single time you use it.

Writing about that every week is good. Demonstrating it every week is better. Marcel now has a desk, a co-host to bully, opinions about luxury knitwear, and a growing list of characters he claims to have had tea with. That’s world-building. That’s the show.

Marcel says: Thanks for watching Character News Network. Don’t sue us.

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - the only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

Nobody agrees on who made it.

It’s a wooden log with a face, holding a baseball bat. It announces the pre-dawn meal during Ramadan. Its name is Tung Tung Tung Sahur, and it might be the most important IP story of the year.

The meme that became a franchise

Quick catch-up if you’ve been spared. “Italian brainrot” is a wave of AI-generated characters that started in early 2025: surreal creatures with pseudo-Italian names, made in minutes, remixed by thousands. A shark in sneakers. A ballerina with a coffee cup head. The log with the bat.

One TikTok user reportedly posted the original Sahur video and pulled 31 million views in a month. Then the internet took over. Thousands of creators gave the character lore, enemies, a universe. No brand book, no committee, no budget. The swarm built a franchise for free.

Then the money showed up, and the whole thing detonated.

According to El País, a European trademark on Tung Tung Tung Sahur has already been granted, with around 20 applications filed worldwide. The US filing sits with a company most fans have never heard of. A Roblox game already had a character pulled over a trademark claim. And the creators of Steal a Brainrot, the Roblox game that aggregated all these characters, are in federal court fighting a company called Mementum Lab over who controls them. Trade press calls it a custody battle. That’s the right phrase. This is a divorce where nobody can prove they were married.

Not gonna lie, I feel very old watching this stuff 😅

Meanwhile, PhatMojo signed a global toy license for Steal a Brainrot. Collectibles hit mass retail later this year.

Read that sequence again. The internet authored a character. Lawyers are assembling the ownership afterwards. And the party capturing the most value isn’t any of the authors. It’s the game that gave the characters a place to play.

Marcel says: authorship is free now. Ownership is what’s left to fight over.

Subscribe now

The mirror image: owning everything, making nothing

Now flip it. The World Cup is the most tightly owned entertainment property on earth. FIFA controls the marks, the stage, and, for 2026, three official mascots, a first: Maple the moose in goal, Zayu the jaguar up front, Clutch the eagle in midfield. Licensed, protected, already in a video game.

And the World Cup content people actually send each other? Fan-made. There’s a full AI-generated football movie on YouTube. Funny Matrix memes. Endless remixes. FIFA owns everything and authors almost none of the culture around its own event. It also failed tremendously in building something worthwhile to watch with its own mascots.

So here are the two failure modes, side by side. Brainrot is authorship without ownership: the crowd makes something priceless it can’t hold onto. The World Cup is ownership without authorship: the institution holds everything, and the crowd makes the magic somewhere else.

Marcel says: a trademark without love is paperwork.

And the pipeline got absurdly fast

This used to take decades. Mickey needed generations to go from cartoon to empire. Tung Tung Tung Sahur went from a TikTok post to trademark filings to a mass-retail toy deal in under two years. Doodles is building custom toys straight from its characters. The distance between “the internet likes this character” and “this character is on a shelf” has collapsed to months.

Which means the ownership question isn’t academic anymore. It arrives with the first viral clip, and it arrives fast.

The playbook

Put the two failure modes together, and the winning position falls out by elimination:

1. Build the character. Real authorship, real love, a personality worth remixing.

2. Own it. Trademark, design registration, the boring paperwork, on day one, not after the third viral moment.

3. Then open the doors. Invite people to play with it. Duolingo lets the internet do unspeakable things to Duo, and the bird only gets stronger. Labubu became a phenomenon because people style it, trade it, and film it. The remixing isn’t a threat to the asset. The remixing is the asset. You just have to be holding the paper while it happens.

Most brands do exactly one of the three. They own a character nobody loves, or they chase a trend they’ll never own, or they build something great and lock it in a brand book where nobody can touch it.

This is the whole Marcel experiment, by the way. Built him, own him, and I hand him to AI tools, to my newsletter, to anyone who wants to play. The internet can’t steal what you’re actively giving away. It can only make it more famous.

Marcel says: build it, own it, then leave the door open.

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - the only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

Don’t. This is the most complete mascot playbook running in Europe right now, and it took eighteen years to build. Follow the career of the guy in the suit, and you can see where every brand character is heading.

Eighteen years of promotions

Billie has been Bol’s mascot since 2008. For most of those years, he did what most mascots do: sit in the corner of the brand, wave, get ignored by the annual report.

Then the promotions started.

First he became the customer service chatbot. The face you complain to. Then, last November, the face of their gift finder. This spring bol announced their store-wide AI shopping assistant, the thing their CEO demos by saying “Hallo Billie, kun je Starbucks capsules voor me bestellen?” Same character. Amazon built the same tool and called it Rufus. There’s a crucial difference: Rufus is a name. Billie is a face people have trusted with their parcel problems for years.

And now, with “Effe bollen,” Billie got the lead role in the brand campaign. A full body, a personality, humour, and a sidekick.

Marcel says: most brands retire their mascot right around year five. bol kept promoting theirs instead.

The campaign: trust, with a face on it

The strategy is sharp. After 27 years, Bol isn’t competing on choice anymore. Their marketing director puts it plainly: in a world where you can buy everything everywhere, the question isn’t where to find it, but who you can trust. “Effe bollen” translates roughly as: relax, it’ll be fine. You can’t shoot a feeling. So they cast two characters who embody it.

The sidekick is the quiet masterstroke. Boxie is built from Bol’s own packaging. The box is the product, the one physical thing this brand ever puts in your hands, and now it has a nose for deals and a face your kids will recognize. Your product dancing through every scene without a single product shot. It never feels like promo, because the promo is busy having a personality.

A duo also solves the hardest mascot problem: dialogue. One character explains, two characters banter. Banter is repeatable. Repeatable becomes a campaign engine. Made by TBWA\NEBOKO, produced by Czar, directed by Floris Kingma, running in the Netherlands and Belgium all summer.

Marcel says: your best mascot might already be in the warehouse.

Subscribe now

Then they printed him

Here’s the part most brands skip. bol didn’t keep Billie on screens.

For the rebrand reveal they had a 3-metre inflatable Billie built, hidden under blue fabric with a countdown timer at the Jaarbeurs. It now stands at their Utrecht headquarters. Four permanent Billie sculptures went up at offices and warehouses. This summer Billie and Boxie are wrapped around the Amsterdam ferry, trams in Ghent and Antwerp, and station walls in Ostend, Ghent and Bruges.

Their CCO calls the duo a distinctive brand asset “recognizable within seconds.” Byron Sharp would nod. A character you can stand next to and photograph is distinctiveness doing physical labour.

Marcel says: a mascot on a screen is an ad. A mascot in the street is an event.

Why this is bigger than Bol

Play the whole arc forward and the pattern snaps into focus.

The character becomes the interface. The box stays the same delivery method. Both have a face. You ask Billie, and Boxie shows up at your door. The entire customer journey, cast as a double act.

That’s the new job description for brand characters. Not decoration. Infrastructure. The voice you talk to, the personality that answers, the shape that shows up in the street and on your doorstep. Every brand building an AI assistant right now is quietly auditioning for this, and most are casting a text box.

Marcel says, " The mascot got promoted to interface. Yours is still in the brand book.

What to steal

Three things, whatever your size:

1. Don’t retire the character; promote it. Billie earned trust in the least glamorous job in the company: customer service. That trust is the entire campaign now.

2. Give your product a face if you can. The box was sitting right there.

3. Print it. An inflatable, a statue, a shelf toy. The physical version makes every digital appearance hit harder.

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - the only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

ChatGPT agrees with you. Claude apologises when you raise your voice. Gemini wants to know how it can help. Every brand sounds the same. Helpful, friendly, slightly desperate.

The work I actually remember this month does the opposite. It builds characters with friction. Weird, gross, mean, joyful in a slightly unhinged way. Impossible to ignore.

Marcel says: “I am chaotic and slightly unhinged. That is the entire point.”

Slather made the sun the bad guy

One line: “The sun is not your friend.”

Every sunscreen brand on earth sells holidays and bikinis. Slather (by SICKDOGWOLFMAN) sells you a creepy, vindictive sun actively trying to cook you. Their latest spot is an Aussie infomercial spoof where the sun BBQs a man on a spit.

Same SPF 50+ that everyone else sells. Suddenly, it’s protection from a threat, not an accessory for a tan. They villainised the entire category’s hero. Men, who barely use sunscreen, finally noticed.

Marcel says: “Make the customer’s problem the antagonist. Always.”

BONUS POINT: I particularly like that this one is embedded fully in the brand; it’s not just a campaign asset.

Four Seasons made STIs into kaiju

A couple skips the condom. Cut to a giant baby kaiju levelling Sydney. Then gonorrhoea shows up. Then chlamydia. Then syphilis. Full grotesque monster movie.

Four Seasons Naked Condoms, by Emotive and AiCandy. No PSA energy. No shame. Just literal cinematic monsters personifying every reason to reach for the box.

The kicker: the kaiju were built to talk back in the comments. Hate the ad? Gonorrhoea pops up to celebrate its case numbers.

“The campaign is built to respond in real time.” — Gavin McLeod, Emotive

A friendly mascot would have killed this on impact. The friction is the point.

Subscribe now

Tesco built a fruit giant called Plumbert 🍎

The masterclass.

Tesco doubled its Free Fruit & Veg for Schools programme. The kind of brief that usually produces a soft-focus ad of kids smiling at apples. Forgettable, every time.

Instead, BBH built a giant. His name is Plumbert. 105,000 photorealistic pieces of fruit and veg. Six months in post. He travels the UK with a boy called Theo, gradually giving himself away to schools until he’s almost gone. All set to “Give a Little Bit.”

It should be cringe. It isn’t. They committed to the character. The agency invented their own term (”fruit resolution”) just to get his apples the right size.

“A big initiative deserves a big idea. Some would say it’s giant.” — Felipe Guimaraes, BBH London

A purpose ad with a narrator dies after one airing. A purpose ad with Plumbert sells plush toys and runs for years.

Marcel says: “Friction isn’t always anger. Sometimes it’s commitment.”

Canva built a squirrel cult and refused to explain it

A giant unbranded squirrel statue appears in Brooklyn Bridge Park. No logo. No press release. Around it: knitting circles in acorn hats, a squirrel choir, balloon vendors, billboards from a fictional movement called The Squirrelites.

New Yorkers had no idea what was going on. That was the whole point.

Canva and Quality Meats sat on the reveal for days. Why a giant squirrel? Who are The Squirrelites? Is this a cult? Eventually, the hero film dropped: a woman has a chance encounter with a squirrel and accidentally starts a movement.

The campaign works because it withholds. Most B2B and SaaS marketing tells you what it is in second one, then explains why you should care for sixty more. Canva did the opposite.

“The squirrel concept works because it is bizarre enough to interrupt attention.” — ContentGrip

Curiosity loops travel faster than clarity. A friendly mascot waving from a Canva banner is invisible. A faintly menacing squirrel that won’t explain itself gets a week of free press.

Admittedly, the character’s versatility also makes it less memorable as a long-term asset, but I’m curious to see if Canva will continue to build this world or if it was a one-off.

Subscribe now

The pattern

Smooth is forgettable. Friendly is forgettable. Helpful is forgettable.

These brands didn’t pick villains. They picked characters with edges. STI kaiju have edges. A creepy sun has edges. A fruit giant who gives himself away has edges. A squirrel cult that won’t explain itself has edges.

Edges create curiosity. Curiosity creates memory. Memory creates brand.

We’ve been here before. The Noid. Mr Mucus. The Pillsbury Doughboy is sixty and still working harder than most social teams. They earned their decades by refusing to be likeable in the bland, agreeable way that AI and modern marketing default to.

Marcel says: “Polite characters die in the algorithm. Distinctive characters live forever.”

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

You walk in. You grab something. You leave. You forget the store existed by the time you reach the parking lot.

That was Miniso, circa 2019. A cheap Chinese lifestyle chain with 2,000 locations and zero reason for anyone to care about it.

Fast forward to today: nearly 8,000 stores across 100+ countries. $2.4 billion in annual revenue. Gross margins of 44.9%. A 100-million-member loyalty program.

The spatulas didn't change. 🛒

The characters did.

What Miniso figured out, and quietly built into a $2.4B machine, is a three-layer IP strategy. Rented characters. Owned characters. Stores are designed as character universes.

Let's break it down.

Tier one: rented IP 🎠

Around 2020, Miniso started licensing characters at scale. Disney. Sanrio. Pokémon. Harry Potter. Barbie. Over 150 global IPs in total, with 10,000+ new IP products launched every year. Annual IP product sales now exceed $1.4 billion.

The blind box category shows this most clearly. In the first three quarters of 2024 alone, they sold more than 30 million blind box pieces. During one New York campaign, blind box sales jumped 64% in a week and generated 170 million social media impressions.

Not from ads. From the characters.

The blind box thing confuses a lot of people: they clock it as a gambling mechanic and move on. But that misses the whole point. Blind boxes only work when you genuinely care which character you get. Nobody queues for a blind box of random generic shapes. The Sanrio collection became Miniso USA's top seller. Disney Villains sold out immediately. US membership grew 250% in 2024. Global membership passed 100 million.

A hundred million people joined a loyalty club for a €5 plastic figure. Sure. The figure just happened to be Hello Kitty. 🐱

Celebrity endorsements work exactly the same way.

Nespresso built a global identity around George Clooney. Instant recognizability, aspirational associations, years of brand equity. Then Clooney aged out of the brief, and they moved to Dua Lipa. Another contract. Another fee. More equity flowing into a celebrity's pocket instead of an owned asset.

That's the ceiling of rented IP. Every impression you build on someone else's character, whether that's Hello Kitty or George Clooney, is equity you're paying rent on. The landlord can raise prices. Or leave.

You can build a very large business this way. But you can't build a moat.

Subscribe now

Tier two: owned IP ✨

Miniso knew the ceiling existed. So they started building their own characters.

PenPen. DunDun. The Gift Bear family. And YOYO, a proprietary character launched in 2025 with a blind box line that sparked immediate fan customization and a wave of user-generated content the moment it dropped.

Projected first-year sales for YOYO: 40 million RMB. From a character they invented. And own outright. Forever.

No renegotiation risk. No licensor walking. The fanbase compounds permanently. Their CMO put it plainly in this article:

A CMO of a $2.4 billion retailer just described their moat as characters and storytelling. Supply chain didn't make the list. Neither did price nor distribution.

The store as a character universe 🏰

The final layer is physical. Miniso didn't just put characters on products. They rebuilt the entire store experience around them.

Their MINISO LAND flagship in Shanghai hit 100 million RMB in sales within nine months. Monthly sales reached 16 million RMB, with 83% coming from IP products. It won Best New Store Concept at the MAPIC Awards 2025.

These aren't stores. They're brand universes with a checkout counter. You don't browse a MINISO LAND. You enter a world. 🌍

When they opened their Jakarta flagship, 3,000 square meters, the largest Miniso store globally, it set a single-day sales record for the entire brand. Jakarta didn't need more stationery. It needed a giant immersive character experience worth traveling to.

Physical activation in an AI-saturated content environment earns disproportionate reach precisely because it's real and impossible to generate with a prompt.

Subscribe now

The playbook 🗺️

Rented IP gets you attention fast. Borrow an existing fanbase, and they do the marketing for you. It works, and Miniso used it brilliantly to scale from nothing to everywhere.

Owned IP is where the business starts compounding. Characters you build belong to you permanently. They grow in value with every product, every activation, every piece of content. YOYO's 40 million RMB debut is what year one looks like when the equity stays yours.

Stores built around those characters are the final layer. They convert a retail visit into a brand memory: something people photograph, share, and talk about long after they've left.

Miniso coined "interest-driven consumption" back in 2020 and quietly assembled a $2.4B company around it.

The question is what you build yours around.

Marcel says: Fascinating case study. Now, someone please explain why YOYO is making 40 million RMB and I'm still waiting on my first royalty check. 👀

Want to build your brand universe?

Book a 20-minute call or browse examples of character-driven brand building in practice.

Not subscribed yet? Join 4,300+ brand builders getting weekly mascot case studies and AI workflows.

Sources

• MINISO Group Full Year 2024 Financial Results, https://ir.miniso.com/2025-03-21-MINISO-Group-Announces-December-Quarter-and-Full-Year-of-2024-Unaudited-Financial-Results

• MINISO Wikipedia, https://en.wikipedia.org/wiki/MINISO

• MINISO USA 250% membership growth press release, https://www.prnewswire.com/news-releases/miniso-usa-achieved-staggering-250-membership-growth-in-2024-underscoring-successful-localization-strategy-302371194.html

• MINISO Blind Box Carnival 20 million pieces, https://www.prnewswire.com/news-releases/minisos-global-blind-box-carnival-spreads-joyful-surprises-as-its-blind-box-sales-top-20-million-pieces-for-the-year-302022785.html

• MINISO proprietary IP incubation plan and YOYO, https://www.miniso.com/brand/news_167.html

• MINISO CMO interview on IP strategy, https://www.brandinginasia.com/qa-robin-liu-ip-led-experiential-retail-is-reshaping-the-foundations-of-global-retail/

• MINISO Super IP and Super Store strategy, https://www.retail4growth.com/news/miniso-scales-up-its-ip-led-super-store-strategy-7503

• MINISO global growth strategy overview, https://matrixbcg.com/blogs/growth-strategy/miniso

Oh, and if you forgot, last month Apple launched a mascot too.

Six stories. One pattern. Let’s go.

1. Pip & Nut: a squirrel with the math to back it up

Pip & Nut is a UK nut butter brand. Their mascot is, predictably, a red squirrel. Nothing clever about the design. It’s a squirrel. It eats nuts.

The campaign animates the squirrel via AI (with creative studio Scary Robots), pairs it with a music choice that does half the emotional work, and scores 5.9 on System1 — top 1% of all TV ads ever measured.

For contrast, McDonald’s recent AI-heavy Christmas adverts were called “creepy” and “uncanny,” and the ad was finally pulled. Same technology. Wildly different result.

The difference isn’t the AI. It’s that Pip & Nut had a character first and used AI to bring it to life. McDonald’s used AI to manufacture a feeling. One is an owned asset getting a new tool. The other is a tool pretending to be an asset.

Marcel says: “Simple character + great music + actually owning the thing = top 1%. Three ingredients, no MBA required.”

2. Decathlon Quechua: a robot walks into a forest

Quechua launched the MH900 hiking shoes with a small white robot called QUECH-01. The setup: QUECH-01 is a product tester at Decathlon’s Mountain Store testing facility who decides the lab tests aren’t enough. So it escapes into the wild to push the shoes harder.

It walks through brambles to test the mesh. Wades through a stream to test the waterproofing. Sprints from a brown bear and does a full Van Damme split between two cliffs to test the grip. Then the human engineers show up, mildly annoyed, and explain that they’re supposed to be the ones testing. Tagline: “Designed to take you further. Tested to go beyond.”

Here’s what makes this work: every outdoor brand on earth says “we test our gear.” Quechua is the only one with a character who does it.

That’s the move. They took the most boring, most over-claimed, most ignored line in outdoor marketing (”we care about quality”) and gave it a backstory. A character. A motivation. Now “we test our gear” is a story instead of a sentence.

And QUECH-01 can come back forever. New shoes? New tests. New tent? Test in a hurricane. The narrative engine is built. They just have to keep feeding it.

Marcel says: “Most brands say ‘we care about quality.’ Quechua built a robot who breaks out of a lab to prove it. Same message. Different universe.”

Subscribe now

3. Indeed’s Ingrid: the recliner that tells you the future is hybrid

If you haven’t watched the making-of for Indeed’s Ingrid character, stop reading this and go watch it. I’ll wait.

Here’s what’s happening: a creature suit, a practical-effects rig sitting in an actual recliner on a studio set, motion capture, and then a digital character layered on top. Old-school puppetry meets new-school post.

Why does this matter? Because the AI mascot conversation keeps getting framed as a binary: real production vs. generated. Ingrid proves the answer is hybrid — use whatever gives you the most soul per scene. Foam puppet, Veo prompt, or both.

The brands that win the next two years won’t be the ones who go “all AI” or “all traditional.” They’ll be the ones who understand that the character is the constant and the tools are interchangeable.

Marcel says: “A robot wearing tracksuit pants in a recliner. This is the most honest behind-the-scenes photo in advertising right now.”

4. Nike Air Max 90: Pablo Rochat builds a character by hand

Pablo Rochat made a character for Nike’s Air Max 90 page that is literally hand-assembled. You can see the strings, the tape, the seams. Deliberately rough, charmingly DIY, wearing the shoes.

This is Nike. They could have generated a perfect 3D character in fifteen minutes with current tools. They chose to build something that looks made.

That choice is the point. In a feed where every product shot looks the same, a character with visible thumbprints stops the scroll. The robot doesn’t need to be technically impressive. It needs to feel like someone cared..

It also signals where brand mascots are heading: not more polish, more texture. The Pip & Nut squirrel works because it feels handmade in the script. The Nike robot works because it feels handmade in the materials. Different routes, same destination.

5. Telenet’s BingeFoot: a Belgian telecom finally gets a face

Belgian telecom Telenet just introduced BingeFoot, and a few sidekicks (including Merbusinessman 🤣). Yes, the name does what you think it does.

For most of the last twenty years, telecom branding has been a race to the same place. Blue. Geometric shapes. Hopeful music. Maybe a couple looking at a phone. Telenet just opted out.

I’ll be watching how this one is applied. The launch is the easy part. The next twelve months of consistent, weird, distinctive use are where most mascots quietly die.

Subscribe now

6. The contrast: George Clooney has to leave room for Dua Lipa

A recent Nespresso-style ad starts with George Clooney, the face of the brand for nearly twenty years. Then Dua Lipa walks in. Clooney has to step aside to make room. It’s played for laughs, but the subtext is brutal: the celebrity you spent two decades building is now sharing the frame with someone half his age, and at some point soon, he’s leaving the frame entirely.

That’s celebrity-led brand building at its endpoint. You spent millions making someone synonymous with your product. Then they age out, get cancelled, or lose relevance, and you pay millions more to bolt the next person on. And again. And again.

Every dollar you spent making Clooney famous for the brand now has to be partly redone for Lipa. The asset doesn’t compound. It depreciates the moment the contract ends.

A character doesn’t have that problem. The Pip & Nut squirrel will still work in 2030. The Decathlon robot doesn’t have a publicist. Pablo Rochat’s Nike character can come back for the Air Max 270, the 95, and the Vapormax. Marcel will still be wearing his stupid orange pants and his MMGA cap when half of LinkedIn has rotated through three different “brand ambassadors.”

That’s the moat. Not aesthetics. Not nostalgia. Asset economics.

So what

Six stories, one pattern: the tool isn’t the strategy, the character is.

If you’ve been waiting for the “right moment” to give your brand a face, the moment was about eighteen months ago. The second-best moment is now, while your competitors are still debating whether ChatGPT is going to replace their copywriter.

Marcel says: “Six brands, one moat. I’d write a thesis, but I’m a dog.”

Want to build your brand universe?

Wonder how your brand universe could look like? Book a 20-minute call here.

Subscribe to Marcel’s Lab for weekly mascot case studies and workflows.

Subscribe now

Prada introduced characters into its advertising

Not a mascot in the traditional sense. Something stranger and more interesting.

For the Spring/Summer 2026 campaign, Prada commissioned artist Jordan Wolfson to reimagine the advertising. His intervention invented unnamed, unreal, dreamlike creatures that interact with the campaign cast in both stills and video. Imagination made tangible, they called it.

These aren’t brand characters with names and personalities. They’re art-world characters: uncanny, unnamed, undefined. But the move itself is significant. Prada, one of the most image-controlled brands in fashion, let invented creatures share the frame with Carey Mulligan and Nicholas Hoult.

The short film builds toward a mantra the cast chants together: “I, I, I, I am...” Left intentionally incomplete.

It’s not a mascot strategy. But it’s evidence that even the most serious brands are reaching for character energy right now. When luxury goes there, something has shifted.

Marcel says: “Unnamed dreamlike creatures that interact with the cast. Finally, my people.”

GifGaff built a billboard you could hug

In the lead-up to what they called “Price Hike Day,” GifGaff partnered with agency Pablo to turn a billboard into a giant, furry, memory-foam embrace. A literal hug on the streets of Covent Garden in London.

There’s a bigger point here beyond the execution. Physical activations used to compete with digital. Now they feed it.

In a content landscape increasingly flooded with AI-generated imagery, a real thing that actually exists in the world stands out immediately. People filmed it, shared it, talked about it. The hug happened once on a street corner and travelled everywhere because it was real, tangible, and slightly absurd. 🫂

This is the new logic of physical activations: they’re not alternatives to social media, they’re fuel for it. And that fuel is getting more valuable, not less. The more AI content normalises, the more a furry billboard with arms beats a rendered visual of a furry billboard with arms. Every time.

Marcel says: “I also give great hugs. I am working on being physically real. Updates to follow.”

Subscribe now

Firefox: two versions of the same mascot

Firefox introduced Kit, its first-ever official mascot, developed with agency JKR under the “More Fire. More Fox.” platform.

It’s quite cute, well-designed. But it’s … a fox. It’s well designed, but it lacks character, story, and behaviour. A stylised, polished launch video by a well-renowned agency doesn’t bring a character to life.

On social, the character falls flat. Sure, they have some plushies, and that’s cute, but there is no strong story.

@firefoxDid someone say more fire, more fox?

Tiktok failed to load.

Enable 3rd party cookies or use another browser

This is one of the most common mascot pitfalls. The brand invests in an expensive launch, the character gets handed to whoever runs social, and the two versions never quite connect. Mozilla described Kit as “a companion, not a commentator, not here to deliver punchlines.” Maybe Kit should actually deliver more than just be a cute, well-designed sidekick.

Compare that to what Apple did, almost by accident.

A few weeks back, I talked about Apple’s Little finder guy. Back then, it was a single mysterious appearance in one of the live feeds on TikTok. Then what happened? The internet blew up over little Finder guy.

Now we see Apple really investing in the character; he even took over the full feed on TikTok.

Alan built a whole universe

Full disclosure: I worked on this one. So take that as you will. But I think it’s worth showing what it looks like when everything actually connects.

For their 10th anniversary, Alan announced a partnership with Kylian Mbappé as ambassador and investor, and unveiled five major health innovations alongside it. But the part that matters most from a brand character perspective is what we built around it: a full brand universe, showing the marmot (yes, it’s a marmot) mascot alongside a new set of characters, including the cute berry (my favourite), all living inside the same coherent world.

Alan’s marmot started as a few curves in Illustrator in 2016. Today, it’s the face of health for over a million members across four countries. That didn’t happen by accident.

The Alanland universe video is that philosophy made visible. The marmot, the berry, new characters, all in the same world, all speaking the same visual language. Product and brand are fully aligned, not running parallel strategies that occasionally acknowledge each other.

That’s what ten years of consistent character investment looks like.

Marcel says: “This is the goal. Also I am available for health insurance partnerships. Very healthy. Very blue.”

Lumo is ugly. That’s actually the least interesting part.

Eurovision 2025 unveiled its mascot, Lumo and the internet had thoughts.

Lumo is a heart-shaped, gender-neutral character designed by Lynn Brunner, a graphic design student from Basel. Despite being entirely human-made, many assumed it was AI-generated.

That’s the first real issue. In an era where AI-generated characters are everywhere, the threshold for “this feels generated” has dropped significantly. A human-made character can now read as AI-made simply because of certain aesthetic choices. That’s a new design problem that didn’t exist two years ago.

The second issue: ugliness is not the problem. Ugly mascots can be great (remember ‘The Noid?’). Weird and unsettling can generate more affection than beautiful and forgettable. The fan content already emerging around Lumo proves it. People are running with the character in ways the official launch never invited them to. 👀

The real problem is that Eurovision approached this without anything resembling a character strategy. They picked Lumo from a student competition and called it a “joyous ambassador of musical diversity and togetherness.” That’s a description, not a personality. No voice, no point of view, no indication of what Lumo does or says or wants. The launch gave people nothing to go on beyond the face, so the face became the whole conversation.

@allthingsadam.ieWhat do you think 😂? #eurovision #eurovision2025

Tiktok failed to load.

Enable 3rd party cookies or use another browser

An ugly character with a strategy can be extraordinary. Lumo doesn’t have one yet. But Eurovision is a long event. How it gets activated in the coming weeks matters more than how it launched.

Marcel says: “Joyous ambassador of musical diversity and togetherness. This is going on my LinkedIn effective immediately.”

Subscribe now

It’s clear: creating a character today is the easy part.

The hard part is making it sticky, interesting, and worth sharing.

Want help figuring that out for your brand? Book a 20-minute call here or see how it works in practice.

And if someone forwarded this to you, subscribe here.

Let’s go.

Apple built a mascot from a 40-year-old icon

Apple joined TikTok. 7.7 million followers. And during a livestream called “Matcha Break with MacBook Neo,” they dropped a tiny 3D figurine based on the Finder app icon.

They called it “Lil’ Finder Guy.” It went viral.

The whole campaign runs on the Gen Z TikTok playbook: nostalgia for a simpler digital era, a slightly absurd physical character, and a format Apple actually committed to instead of parachuting in a polished ad.

The best part? They did not design a new mascot from scratch. The character was sitting inside their OS for 40 years. They just gave it a body.

Now the only question is, was it a one-off stunt or will we see more? Give the people what they want, Apple!

Marcel says: “A little blue face on every Mac since 1984 and nobody gave it legs until now. Forty years of free mascot, just sitting there.”

The puppet data came in and it is embarrassing for everyone

Andrew Tindall analysed 1,265+ global ad campaigns representing $140 billion in market share. System1 tested every ad on 250,000+ people. Here is what they found:

Puppet and animated character use dropped from 26% of campaigns in 2012/2013 to just 5% in 2022/2023.

While the ROI data kept getting better.

Ads with puppets or animated characters are +25% more emotional on average. They deliver a +5pp pump in brand recognition without the celebrity spillover problem. And they produce +24% more business results including profit, share growth, and reduced price sensitivity.

The reason he posted all of this? A bear.

GUT D3 and Puppet Island created FiBEAR — a warm fuzzy puppet bear — to launch Smartfood FiberPop for PepsiCo.

A great puppet character makes you care about a category you had zero interest in. That is not a creative trick. That is the business case.

Marcel says: “Puppet use dropped 80% over ten years while the data kept screaming USE MORE PUPPETS. A bear named FiBEAR is making the argument better than any deck ever could.”

Subscribe now

Penguin freed a 90-year-old bird

Penguin Random House UK just introduced “Playful Penguins”: a suite of hand-drawn illustrations that takes their iconic logo bird and gives it room to move.

The original penguin was sketched in 1935 after a trip to the London Zoo. For the last decade it has mostly lived inside an orange oval. Now it jumps, reads, dances, and slides across the page.

The logo stays. The core mark is untouched. But now the character has a whole system to live in: seasonal campaigns, social, point of sale, social impact work. Less corporate mark, more cast member.

This is the difference between a logo and a mascot. A logo is a mark. A mascot can go places the logo cannot. Penguin had the character the whole time.

Marcel says: “Ninety years of brand equity and they just now taught the bird to dance. Late. But not too late.”

3 brands. 3 different ways to answer the same question: how do you become impossible to ignore when everything is noise?

You build something people remember. A tiny guy from inside your OS. A bear who cares deeply about fiber. A penguin that finally learned to move. When content is infinite, characters are your moat.

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe now

And both will be forgotten by March.

Not because they were bad. They were genuinely entertaining. But because they were built on rented territory. Xfinity doesn’t own Jurassic Park (well, Comcast owns Universal, but you get my point). Dunkin’ doesn’t own Friends or Seinfeld or Good Will Hunting. The moment those licensing deals end, the nostalgia walks out the door with them.

Now compare that to what Kellogg’s did with Tony the Tiger.

The tiger that lives on Spotify now

Frosted Flakes partnered with Grammy-nominated rapper JID to remix their ‘90s “Hey Tony” jingle into a full-length track. It’s on Spotify. On Apple Music. In people’s playlists next to actual music, not sandwiched between commercial breaks.

JID grew up eating Frosted Flakes. The lyrics reference his childhood. It’s nostalgic, sure. But it’s nostalgic through a character that Kellogg’s has owned since 1952.

And this is just the latest move in Tony’s multi-front comeback. He became a VTuber on Twitch in 2022. He’s on TikTok. He sponsors the Sun Bowl (five consecutive years). He reunited with Tony Hawk for the Pro Skater 3+4 remake, complete with “Tony x Tony” cereal packs. That’s not a campaign. That’s a character operating across streaming, gaming, live events, merch, and culture simultaneously.

Dunkin’ spent millions assembling ‘90s sitcom stars for 60 seconds of airtime. Kellogg’s put a rapper on a 73-year-old tiger’s jingle and now owns a track on Spotify forever.

Marcel says: “One of these is a firework. The other is compound interest. I know which one I’d rather be.”

Renting nostalgia vs. owning it

This isn’t a knock on Xfinity or Dunkin’. Their Super Bowl spots were objectively fun. The Dunkin’ ad had people rewatching it frame by frame to catch all the sitcom references. The Xfinity spot generated massive conversation, even if most of it was about the uncanny valley de-aging (some critics compared it to a death mask).

But there’s a structural difference between what these brands did and what the mascot brands are doing right now.

When you build a nostalgia campaign around licensed IP or celebrity casting, you’re renting emotion. The warm feeling belongs to Jurassic Park, not Xfinity. To Good Will Hunting and Friends, not Dunkin’. When the campaign ends, you start from zero again next year. Dunkin’ knows this, which is why Affleck is now in his fourth consecutive Super Bowl spot for them. They’re essentially trying to turn a celebrity into a recurring character because they don’t have one.

Now, some brands can lean on their own heritage. Harley Davidson, Coca Cola or Guiness don’t need to license any nostalgic IP, they have their own nostalgia triggers. But those are brands with decades (sometimes over a century) of cultural presence baked in. That’s the exception, not the playbook. Most brands aren’t old enough or iconic enough to be their own nostalgia trigger.

Characters shortcut that problem. When you build a nostalgia campaign around an owned character, you’re depositing into an account that compounds. Every activation adds another layer of memory. Every remix, every collab, every merch drop builds on top of decades of existing mental availability.

The data backs this up. Kantar found that nostalgic ads increase enjoyability by 15 points and distinctiveness by 14 points. But here’s what matters for the ownership argument: characters already outperform celebrities in long-term brand equity. Stack nostalgia on top of an owned character, and you’re combining two of the most powerful forces in marketing.

The mascots you thought were dead just became the most valuable assets in marketing

Tony isn’t an isolated case. 2025 might go down as the year brands collectively realized they’d been sitting on gold mines.

Cornelius the Cockerel

Kellogg’s dropped £12 million on upgrading their 66-year-old rooster from flat illustration to full 3D animation by Framestore (Paddington, Harry Potter). System1 scored the ad at 3.8 stars, above the cereal category average. The brand admitted they’d “lost their way.” Their fix wasn’t a new campaign or celebrity deal. It was investing in the character they’d had since 1957.

The payoff? Cornelius got promoted from Corn Flakes mascot to masterbrand character across all Kellogg’s cereals.

Pillsbury Doughboy

AdWeek’s standout brand mascot of 2025. At 60 years old, he got a modern makeover, a 25-foot inflatable in Brooklyn, an AR “Doughhouse” experience, a Jimmy Fallon collab, and a reissue of the 1988 cookie jar that sold out in 24 hours.

General Mills reported more social engagement in Q1 than in all of fiscal 2025. Their VP called the Doughboy “Pillsbury’s strongest asset.” Not their recipes. Not their distribution. Their character.

Marcel says: “A 60-year-old piece of dough with a belly you can poke is outperforming entire social media teams. Think about that.”

And they’re not alone

Tootsie Roll refreshed the “How Many Licks?” commercial for the first time in 55 years, redrawn in HD, same storyline, same Mr. Owl. Cheez-It un-retired Prince Cheddward. Domino’s launched Mac Scott and gained 10,000 followers in a day. Aldi UK’s Kevin the Carrot married his sweetheart Katie after nearly a decade of storyline. And the Instacart Super Bowl spot crammed the Green Giant, Chester Cheetah, Kool-Aid Man, Pillsbury Doughboy, Energizer Bunny, and Mr. Clean into one ad, because every one of those characters already lives in our heads rent-free.

Everywhere you look, characters are coming off the bench.

Why now?

Nostalgia is the dominant emotional strategy in marketing right now, and the numbers explain why.

Nielsen says 61% of millennials say nostalgia improves their brand perception. 68% of Gen Z responds positively to throwback marketing even for eras they never lived through. In uncertain times, people reach for the familiar.

But nostalgia alone doesn’t explain why everyone’s specifically dusting off their mascots. Two things do.

• Control. When you build campaigns around celebrities or licensed IP, you’re at the mercy of someone else’s reputation, availability, and price tag. Characters are your IP. You control the narrative, the scheduling, the evolution, and the cost. No contract renegotiations. No reputational risk you didn’t choose. No competing endorsements. That level of creative control becomes increasingly valuable when the cultural landscape is unpredictable.

• AI is making distinctiveness the only moat. When anyone can generate a month’s worth of social content in an afternoon, the brands with recognizable, owned characters cut through the noise at thumbnail size. The same AI abundance that makes production cheap makes established characters more valuable, not less. Characters are up to 15x more memorable than logos. They consistently outperform celebrities in brand recognition and correct attribution. And they compound over time instead of expiring after a campaign cycle.

Marcel says: “I’ve been saying this for two years and people thought I was biased because I’m a mascot. Turns out all these billion-dollar companies agree with me.”

The playbook

Looking across these revivals, there’s a shared pattern.

• Visual upgrade, personality preserved. Cornelius went from flat to 3D. The Doughboy got a wardrobe refresh. Mr. Owl got HD. In every case, they changed the resolution, not the character.

• Culture collaborations rooted in authenticity. Tony x JID works because JID genuinely grew up with Frosted Flakes. Tony x Tony Hawk works because they share ‘90s commercial history. These aren’t random celebrity pairings.

• Platform expansion beyond advertising. The JID track lives on Spotify, not in a TV spot. Tony is a VTuber. The Doughboy has an AR experience. The best revivals put the character where audiences already are.

• Physical artifacts that extend the memory. The Pillsbury cookie jar sold out in 24 hours. Kellogg’s put “The OG” on billboards. Physical touchpoints extend the character beyond screens and into people’s homes.

What this means for you

If you have a mascot collecting dust somewhere in your brand guidelines, this is your wake-up call. You’re sitting on an asset that’s worth more now than when you created it. The brands winning aren’t inventing new characters (though some are). They’re investing in the ones they already have.

If you don’t have a character yet? Every brand that revived a mascot this year wished they’d never shelved it. Kellogg’s literally said they “lost their way.” Pillsbury called the Doughboy their “strongest asset” after years of underinvestment.

The best time to build a character was 20 years ago. The second best time is now.

Celebrity nostalgia is a firework. Character nostalgia is compound interest.

Choose wisely.

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

→ Book a 20-minute call | See examples

If you’re not subscribed yet, join 4,300+ marketers getting weekly mascot case studies and frameworks:

Subscribe now

IKEA hired the Tiny Chef

The Tiny Chef, a 6.5-inch tall stop-motion creature that looks like a sentient pea in a chef hat, got cancelled by Nickelodeon in mid-2025 after winning a Children’s Emmy. The creators put out an SOS. Then IKEA showed up.

Three episodes follow Tiny Chef walking into an IKEA store looking for a spatula, finding a job application, and becoming an ambassador for their new falafel balls. It works because IKEA didn’t try to change what makes Tiny Chef special. They just gave him a new stage.

The character-brand fit is perfect. Tiny Chef is vegan, cooks plant-based meals, and radiates warmth. IKEA is pushing plant-rich eating. This isn’t a celebrity endorsement. It’s a brand adopting a character whose values already align with theirs.

The handmade stop-motion aesthetic stands out in a sea of slick AI content. Tiny Chef’s mumbling voice is instant sonic branding. Sometimes the most future-proof move is to go analogue.

Marcel says: “IKEA basically did a rescue adoption of a mascot. That’s not a campaign, that’s a brand universe crossover episode.”

Subscribe now

Brita built an entire universe out of a water filter

@britausai'm glowing, i know

Tiktok failed to load.

Enable 3rd party cookies or use another browser

A water filter brand has no business being this entertaining. Brita USA on TikTok has built a full brand universe: recurring characters, original songs, sonic branding, and a visual identity that’s hypnotizing.

The star is the Brita Shark, a green animated shark who sings hydration anthems like “At Least I’m Hydrated” (over a million likes on that one alone). But there’s also Brita Bot doing emo monologues about dehydration, a kawaii kitten, and a rotating cast that keeps the tone consistently weird.

The universe is complete. Visual consistency across characters and animations. Sonic branding with original songs people actually stream on Spotify. A tagline that became a lifestyle. And fan engagement where people draw fanart of a water filter mascot. 376K followers and 11.9M likes on TikTok. For a water filter.

Marcel says: “You don’t need to be a cool category to build a cool character. You just need to commit to being weird and never apologize for it.”

Jet-Puffed’s masochistic marshmallows

A marshmallow brand that hadn’t run a creative platform in almost 70 years just launched one. And it’s about marshmallows who love being tortured.

Kraft Heinz and GUT Miami created “Love ‘Em to Death” for Jet-Puffed, built around the insight that marshmallows exist to be stretched, squished, melted, and roasted alive. The characters are adorable and cheerful about their painful destiny. Unsettling and impossible to look away from.

They teased it with a Las Vegas Sphere takeover. But what interests me most is the collectibility angle. These characters have the kind of weird, sticky (pun intended) personality that makes you want a plushie on your desk. When more than half of people can’t name a marshmallow brand, you don’t need to be polished. You need friction.

Sweet characters meeting joyful demise. Dark humour in bright colours. Don’t give people what they expect. Give them something they can’t forget.

Marcel says: “Todd Kaplan moved from Pepsi to Kraft Heinz and immediately greenlit characters with a death wish. Respect.”

Shareman makes property tech feel human

This one’s quieter. Hymn, a Swiss design studio, created the identity for Shareman, a property tech platform by eeproperty that handles shared services like laundry rooms, EV charging, and solar energy.

The centrepiece is a simple hand character that acts as a “benevolent companion” through the digital experience. It guides users, turns instructions into conversations, and makes interactions feel warm instead of transactional. Three sub-brands (Wash, Charge, Shine) each feel distinct while unmistakably belonging to the same family.

No abstract building silhouettes. No cold minimalism. Just a character that’s calm, friendly, and works across every touchpoint, from smartphone screens to basement laundry room signage. Simple without being simplistic.

Hat tip to from Good Copy for flagging this one (and for all the other mascot gossip).

Marcel says: “This is proof that characters aren’t just for consumer brands going viral on TikTok. Even B2B property tech can feel human with the right mascot.”

Subscribe now

Bonus: Marty Supreme is a brand building masterclass disguised as a movie campaign

This one isn’t a mascot. But the campaign A24 ran for Marty Supreme is the best case study in distinctive brand assets I’ve seen in years.

Marty Supreme: Josh Safdie’s ping pong drama starring Timothée Chalamet. An indie sports movie about a 1950s table tennis hustler. On paper, a nightmare to market. In reality, A24’s highest-grossing film ever at $149 million worldwide, nine Oscar nominations, and a cultural footprint most billion-dollar franchises would envy.

Distinctive codes, everywhere. They picked a corroded, rusted orange inspired by ping pong balls and committed completely. Chalamet pitched it in a fake leaked Zoom meeting, explicitly referencing how Barbie owned pink. Then they did it. Orange blimp over LA. The Sphere turned into a giant orange ping pong ball. Ping pong balls raining at Camp Flog Gnaw. Orange Wheaties boxes. One colour, relentlessly reinforced.

A character that owned it. Chalamet didn’t promote the movie. He became Marty. Talk show appearances flanked by suited men wearing giant ping pong balls as heads. Standing on top of the Sphere declaring “Marty Supreme, Christmas Day.” Cranking Soulja Boy in Brazil while volleying ping pong balls into the crowd. The character bled into real life so completely fans couldn’t tell where Chalamet ended and Marty began.

Viral fashion as brand asset. The $250 windbreaker became a cultural object. Seeded to Kendall Jenner, Kid Cudi, Tom Brady, Bill Nye, Ringo Starr. Pop-ups drew crowds so big the LAPD showed up. Merchandise as a distinctive code, not just revenue.

One message, repeated. “Marty Supreme. Christmas Day.” Every stunt, every appearance, every post. Byron Sharp 101: build mental availability through consistent, repeated codes. A24 did it for a movie better than most brands do it for products with unlimited shelf life.

The result? TikTok tributes still popping up (like this one from @kidondrej). Branded ping pong balls selling on eBay. The jacket is a collector’s item. A niche indie film about ping pong became inescapable.

If Chalamet ever gets tired of acting, he has a future in brand strategy.

Marcel says: “A24 turned an indie ping pong movie into the most distinctive brand launch of 2025. Meanwhile, most actual brands can’t even commit to a colour.”

Characters are everywhere if you’re paying attention

From sadistic marshmallows to a rescued puppet chef, from a singing shark to a helpful hand, and a ping pong movie that outbranded most actual brands, these are all very different executions. But they prove the same thing: characters and distinctive codes turn forgettable into unforgettable.

Whether you’re a 70-year-old marshmallow brand launching its first campaign, a furniture giant looking for the perfect ambassador, a water filter company that wants to be entertainment-first, or a property tech platform that wants to feel human, characters are your answer.

It’s a great time to be a mascot (builder).

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

The Olympics, the Super Bowl, these should be moments where creativity is on full display; instead, we got this:

Cheap. Uninspiring. Unmemorable.

And then this during the Winter Olympics opening:

In some cases …. borderline unethical.

If this is the future of brand building, it’s bleak.

So here’s my plea: we can do better.

Not (just) because it’s noble. But because actual entertainment drives attention (aka profit). Distinctive assets and characters compound. You’re leaving millions on the table using AI as a replacement for creativity instead of a creative multiplier.

What AI should actually be doing

A kid draws. Crayon on paper. Pure imagination

Then someone brings it to life with AI. Amplifying what was already there.

That’s what AI should be doing for your brand. Taking the raw, distinctive, actually interesting ideas your team has and giving them superpowers.

Not replacing the vision. Amplifying it.

Subscribe now

When brands actually give a shit

BBC Winter Olympics 2026. 700 individually 3D-printed athletes. 14 controlled combustion techniques. Real fire. Stop-motion animation. Their creative director said it best: “I hope the industry will see the value and importance of keeping real craft in advertising.”

Zero AI. 700 handmade figurines. Real fire. Pure craft that makes you feel something.

Or what about this Super Bowl ad? Nostalgia, celebrities, and long-term brand-building all wrapped up in an epic sitcom-style ad. Was AI used? Yes, to de-age the actors (similar to how Xfinnity did in their recent ad). Was it used to somehow replace creativity? Definitely not.

Notion’s AI companion. Hand-animated frame by frame. Every gesture is intentional. Then AI made it conversational and scalable. It feels human because humans made it first.

Notice the pattern? The BBC used zero AI. Dunkin’ used AI to enhance. Notion blended craft with AI. And now, one creator used 100% AI video to build something genuinely great:

Lil Wukong. Chinese mythology meets hip-hop meets AI video. 500k views. Not because it’s AI. Because it’s good.

It was never about AI or no AI. It’s about whether someone had a creative vision worth following. A massive broadcaster, a global brand, a tech company, and a solo creator all made something people actually want to watch. Different budgets. Different tools. Different amounts of AI. Same ingredient: they gave a shit.

Skip that ingredient, and no amount of AI will save you.

Subscribe now

Want to build a brand mascot that doesn’t make you cringe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

READY? FIGHT!

The debate is as old as advertising itself: do you borrow fame or build it?

Celebrities bring audiences. Characters bring control. Both promise to make your brand memorable.

But which actually wins?

Let’s settle this the only way that makes sense: a five-round fight to the finish.

In the red corner: The Celebrity – famous, expensive, comes with baggage (literal and metaphorical).

In the blue corner: The Character – created, controllable, takes time to build.

ROUND 1... FIGHT!

Round 1: Fame

Winner: Celebrity 🏆

Let’s give credit where it’s due.

Take this year’s Super Bowl. One ad rolled out Jon Hamm, Scarlett Johansson, and a roster of A-list talent. Millions of views. Instant cultural moment. Social feeds are exploding.

That’s the power of borrowed fame. You don’t need to explain who these people are. They come pre-loaded with attention.

A new character? Nobody knows them. Nobody cares yet. You’re starting from zero, and building recognition takes years.

If you need attention right now, celebrities deliver. That’s not even close.

But here’s the asterisk: those talent fees? Probably north of $10 million. And next month, those same celebrities will be selling something else for someone else. The brand is left with... a memory. A moment.

Nothing owned. Nothing compounding.

Celebrity takes Round 1. Enjoy it.

Round 2: Distinctiveness

Winner: Character 🏆

Here’s where things get interesting.

Yes, celebrities are famous. But they’re famous to everyone. Scarlett Johansson has done campaigns for dozens of brands. When you see her, do you think of any single one?

Now think about Tony the Tiger. The Michelin Man. The Duolingo owl.

Those characters ARE the brand. They’re not borrowed distinctiveness – they’re owned distinctiveness. No competitor can license them. No other category can dilute them.

A 2024 Kantar report found that characters consistently outperformed celebrities in long-term brand equity scores across global markets. The reason? They’re singular. Unmistakable. Yours.

Character takes Round 2.

Round 3: Entertainment

Winner: Character 🏆 (close fight)

This one’s tighter than it looks.

Celebrities can be genuinely entertaining. Ryan Reynolds turned Aviation Gin into a content machine. Matthew McConaughey made Lincoln ads weirdly compelling.

But here’s the difference: those celebrities have boundaries. Schedules. Opinions about what they will and won’t do.

Characters? You can make them do anything.

The Duolingo owl can threaten users on TikTok. The Scrubbing Bubbles can go to therapy. The M&M’s can have complicated relationship drama.

You can take creative risks that no celebrity’s PR team would ever approve. You can be weird, experimental, and prolific – because your character shows up whenever you need them.

The creative ceiling is just higher.

Character takes Round 3. Barely.

Round 4: Playability

Winner: Character 🏆

Want Timothée Chalamet to sign autographs at your brand activation? Get in line. Behind the movie studios, the fashion houses, the fragrance brands, and everyone else fighting for a slice of his time.

Celebrity access is brutally competitive. They have endless options, and you’re renting a tiny piece of someone else’s universe.

But when you own a character, you own a universe.

Merchandise. People buy plushies of the Geico gecko. Nobody’s buying plushies of your celebrity spokesperson.

Games. Characters can star in branded games, apps, and interactive experiences.

Fandom. Strong characters generate fan art, cosplay, remixes, UGC – all of it spreading your brand for free.

Collaborations. Your character can cross over with other brands, other IPs, other worlds.

When you build a character people love, you’re not running campaigns. You’re building a platform.

Character dominates Round 4.

Round 5: Risk

Winner: Character 🏆🏆🏆 FATALITY

This is where the fight ends.

Every celebrity comes with a hidden risk profile. Their past. Their present. Their 3am tweets. Their “private” behavior might not stay private.

(Somewhere, a brand manager is still having nightmares about those flight logs.)

When a celebrity implodes, your brand is collateral damage. You scramble to pull ads, scrub assets, and issue statements. Years of association, destroyed overnight by something completely outside your control.

Characters don’t have this problem.

Your character will never:

• Get arrested

• Say something unhinged on a podcast

• Show up in a scandal

• Demand renegotiation after going viral

• Die unexpectedly

...unless you want them to.

That’s the point. Duolingo killed Duo and got billions of impressions. Liquid Death’s mascot is literally a murderous cartoon. Scrub Daddy flirts with chaos daily.

The difference? Controlled risk vs uncontrolled risk.

You can make your character controversial, edgy, unhinged – even “kill” them for engagement. But it’s YOUR decision. YOUR timing. YOUR narrative.

With celebrities, the chaos chooses you.

Character wins by knockout.

Subscribe now

Final score: Character 4, Celebrity 1

Look, celebrities aren’t useless. If you need instant fame, they deliver. If you have unique access for some reason, USE IT.

But fame is round one. The rest of the fight – distinctiveness, entertainment, playability, risk – all goes to characters.

And here’s what the scorecard doesn’t show: the compounding effect.

Every dollar you spend on a celebrity is rent. The contract ends, you start over.

Every dollar you spend building a character is equity. The recognition compounds. The creative library grows. The brand universe expands.

Over five years, over ten years, over thirty years... it’s not even close.

The exceptions

Some celebrity partnerships genuinely work:

• George Clooney × Nespresso – 18 years of consistency (the exception, not the rule)

• Ryan Reynolds × Aviation Gin – he owned equity, then sold for $610M

• Rihanna × Fenty Beauty – she IS the brand, not a spokesperson for it

Notice the pattern? The partnerships that last are the ones where the celebrity basically becomes a character – either through unusual longevity or actual ownership.

If you’re not getting that kind of deal, you’re probably in a rental agreement disguised as a partnership.

The bottom line

Unless you have unique access to talent that nobody else can get, the math favors building your own characters.

Not because celebrities don’t work short-term. They do.

But because you’re not building for the next campaign. You’re building a brand universe that compounds for decades.

Borrow fame, or build it.

Only one of those shows up on your balance sheet.

GAME OVER. CHARACTER WINS.

Subscribe now

Want to build your brand mascot?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Stranger Things wrapped its final season.

But the universe didn’t end.

Now they can create spinoff series. Build theme park experiences. Launch video games. Expand into new storylines with different characters.

Why? Because they didn’t just make a TV show. They built IP. A universe rich enough to tap into for years.

Think about what made that possible. Not a style guide. Not brand guidelines. A world with specific time and place, recognizable characters with depth, consistent mood and physics, conflicts that extend beyond one storyline.

Every new extension taps into that established world without starting over. The IP compounds. Remember this bunch? Nintendo’s universe can unfold across decades of movies, media (gaming, film, comics, …), while still retaining its distinct ‘Nintendo’ feel.

Most brands will never do this. Not because they can’t, but because they think a brand book is enough.

It’s not.

You don’t need Hollywood money

Marcel’s Lab (this newsletter) runs on a one-person business model. The Marcel universe isn’t massive: a few thousand followers, a couple of hundred true fans.

But it’s real.

Marcel is a slightly unhinged, celebrity-obsessed blue dog who does dumb shit trying to hack his way into fame. He’s clumsy. He acts like he’s real even though he’s obviously not. There’s a fun meta layer: am I the puppet master, or is he?

That character becomes a portal into the universe.
As always, building this universe follows this playbook:

1. Level 1 — Distinctiveness: People first have to notice your character (do the test if you’re curious)

2. Level 2 — Entertainment: Then they get entertained by it (hopefully).

3. Level 3 — Play: Then they interact with Marcel on the Sandbox (level 3), where they’ve generated hundreds of images playing in his world.

Not millions of impressions. Hundreds of fan creations. That’s people creating within your universe, not just consuming your content.

The universe keeps growing. It’s not just a “weird blue dog” .png slapped on social media posts or packaging. It’s how he engages with other mascots, with the audience, the dynamics, the lore that builds over time.

This is brand IP at the scrappy, one-person scale. Tailored to what I need, not some cinematic universe.

Marcel: “If I can build a universe with zero budget and questionable Photoshop skills, your brand has no excuse.”

Subscribe now

The problem: guidelines aren’t IP

Your brand guidelines are necessary. Colors, fonts, and logo usage rules. Nobody’s arguing against that.

But here’s what most brands do: hand an agency (or in-house team) a request with some business objectives (hopefully), a key message (unlikely), and a link to some brand guidelines and say “go create something.”

The brand book tells them how the graphic design should look. The brief tells them what the client expects. It doesn’t give them a world to build on.

So every campaign starts from zero. Every quarter brings disconnected creative. Nothing compounds into something memorable because there’s no throughline, no universe to keep tapping into.

Compare the Market faced exactly this in 2009. They were fourth in a category of four UK price comparison websites. Generic. Forgettable. Insurance advertising is at its most boring.

Then someone pitched something ridiculous: “You’re CompareTheMarket.com, but what if we make ads for a different website? CompareTheMeerkat.com.”

They created Aleksandr Orlov, a Russian aristocrat meerkat frustrated that people keep confusing his site with theirs. His catchphrase “simples” entered the MacMillan English Dictionary. The fictional village of Meerkovo became real to audiences. Baby Oleg joined the family. Autosergei. Meerkat Movies offered cinema tickets. Toys became collectibles.

Within a year, Compare the Market jumped from #4 to #1 in their category. Now they hold 54.5% market share.

Not because they had better guidelines. Because they built a world people wanted to engage with.

That’s the difference between rules and IP.

Building worlds, not campaigns

Media companies understand this instinctively. They’re not building campaigns. They’re building IP. Recognizable worlds they can explore for decades.

Brands? They’re stuck in short-termism. Optimizing for this quarter’s metrics. Terrified to invest in anything that won’t show ROI by Friday.

But look at what happens when brands think like IP owners instead of campaign managers.

Salesforce sells complex B2B software. They could’ve gone the traditional enterprise route: serious, formal, corporate. Instead, they said

Starting in 2014 with Astro the astronaut, they built an entire character universe. Nearly 20 characters now. Codey the bear for developers. Ruth the elephant for architects. Cloudy the goat for admins. Each represents a different role in their ecosystem.

For a B2B brand selling enterprise software, this was radical.

It worked. Their numbers nearly doubled after adopting the character universe. Why? Because boring software becomes memorable when there’s a world to explore.

Progressive Insurance did the same thing in 2016. Homeowners insurance. Not exactly thrilling territory. But they didn’t create an insurance ad. They created a universe around “Parentamorphosis.” The affliction where young homeowners start turning into their parents.

Dr. Rick, the “Parenta-Life Coach,” helps them avoid disaster. Weather obsession. Socks with sandals. Becoming the neighborhood mayor.

Eight years later, 18+ commercials deep, it’s still running. Dr. Rick published a book. He appeared on Hot Ones. He has airport vending machines. The concept entered cultural vocabulary.

That’s not a campaign. That’s IP compounding.

Marcel: “You can’t build a universe when your CMO changes strategy every six months. But that’s exactly why the brands that DO commit to it absolutely dominate.”

Subscribe now

Characters are shortcuts, not requirements

Characters work because faces create instant portals. We’re wired to wonder: Why does this character behave this way? What motivates them? What world do they inhabit?

That’s why Compare the Market’s meerkats worked. Why Salesforce’s Astro works. Why Dr. Rick works.

But fictional characters aren’t the only path.

Ryan Reynolds built an entire universe around Ryan Reynolds. Aviation Gin. Mint Mobile. Wrexham AFC. Every ad he touches feels like it exists in the same world. The world where Ryan Reynolds is a charming, self-aware smartass who understands marketing is ridiculous and leans into it.

That personality creates universe coherence. No fictional character (except for Mint Mobile’s mascot) is required when the founder’s persona is strong enough to anchor everything.

MrBeast does the same. Game show competitions. Feastables chocolate. Comic books. Everything feels like MrBeast because everything operates on the same physics: absurd scale, entertainment-first, generosity as spectacle.

Most brands can’t do this. Most founders aren’t Ryan Reynolds or Jimmy Donaldson.

Which is exactly why most brands need characters. Because most brands don’t have personalities strong enough (or celebrities) to carry a universe alone.

Gen-AI plays nice with universes (and characters)/

Generative AI is terrible at rigid guidelines and incredible at synthetic universes.

Feed AI: “Use Pantone 286C, include our logo, maintain 20% white space.” Result: It hallucinates. Breaks your rules. While at the same timing generating generic slop beneath because you didn’t feed it rich context.

Feed AI: A rich, multimodal universe (images, video, text, audio that all feel part of the same world). Result: Varied output that stays coherent because it’s drawing from deep context.

The brands that will scale AI-native content aren’t the ones with perfect guidelines. They’re the ones with universes rich enough that AI can hallucinate within them productively.

This isn’t five years away. It’s happening now.

What is your favorite brand universe? I’m curious!

Subscribe now

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Carrefour’s “Pricebiter” creature

A furry white creature biting into price tags. Consequently, his shape is also the negative shape of Carrefour’s logo. That’s it. That’s the whole concept. And it’s absolutely perfect.

“Hap uit de prijs” translates to “bite out of the price,” which is probably the most literal mascot execution I’ve seen all year. But it’s also simple enough that it works. It makes ‘discount’ feel a bit less cheap and a bit more fun. And I think that’s smart!

Marcel says: “Notice how they didn’t try to make it cute in a safe way. It’s slightly unsettling. That’s the good stuff.”

Coincidentally, Delhaize also revamped its ‘little’ lions, cute, fluffy versions of the iconic Lion logo. I guess everyone is doing cute discount characters now!

Have a look at some of the amazing behind-the-scenes here in Volstok’s post.

Zyla’s sloth for an energy drink

An energy drink with a sloth mascot. Let that sink in. The entire brand positioning is built on contrast, which is either genius or completely unhinged. I’m leaning toward genius.

There is also the interesting choice to have an illustrated character on the pack, making it very recognizable, and then using the sloth suit (admittedly a lot more generic) on social media and in ads. It makes it easy for the team to make fun content, it feels authentic and a bit weird, but that’s exactly why Duo’s suit also works on TikTok.

This is what I mean when I say characters create friction. Nobody asked for a sleepy animal to sell you energy, but now that it exists, you can’t forget it. It’s contrast, it’s memorable.

Anthpo’s talking iPod Nano

This guy gets it. Anthpo turned an iPod Nano into a full personality with a mission to shake up the music industry. He could have used his face, but he knew that to reach a broad audience, a simple, cute talking character would do a better job.

And this isn’t even his first mascot rodeo. He previously did a bit with a kid wearing Crocs as a character. He understands that objects + personality = infinite content potential.

Marcel says: “This is playability at work. Once you give something a face and a voice, you can deploy it anywhere. Commentary, reactions, announcements, whatever.”

The burnt peanut’s unhinged live avatar

A burnt peanut. With a live avatar. Being completely chaotic on stream.

“The Burnt Peanut - now with millions of YT and Twitch subscribers - is one of the fastest growing and most watched streamers in the world. Here's his fascinating story.”
Full post by John James on LinkedIn

This one breaks my brain in the best way. Unhinged VTuber uses a live peanut skin to go full berserk on YouTube.

What’s interesting to me is how he can be much more of a wild personality on camera behind his ‘peanut mask’. It’s also fun to see how there is almost a ‘tribal’ usage of words like ‘Goop’ with double meanings to signal that you’re part of the ‘bungulator fam’.

What I also love is how the Burnt peanut appears on shows as the avatar; he never appears as ‘the human behind it’, he stays in character, even at an award show he won.

Marcel says: “The production quality sucks. Nobody cares. The personality is doing all the heavy lifting.”

It’s a great time to be a mascot

From cute little red lions to an unhinged, cursing Vtuber Peanut, these are great times for the character renaissance.

They all prove the same point: characters aren’t decoration. They are entertainment. They’re your moat in 2026.

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

Most brands are avoiding at all costs to be anything but perfect, friendly, and nice.

Meanwhile, some of the most iconic characters and campaigns in advertising history are literal villains who represent the exact problem you’re trying to solve.

It’s time for sabotage.

The Noid (Domino’s, 1986-1995, revived 2021)

A buck-toothed gremlin whose job is destroying your pizza. He freezes it, squashes it, and makes it late. Domino’s entire campaign was “Avoid the Noid.”

They personified delivery disaster. Made him cute. Put him in commercials. And, of course, you know the playbook by now: distinctiveness, entertainment, and playability.

Result: One of the most recognized ad campaigns of the 80s. Brought back 30 years later because people still remember him.

Mr. Mucus (Mucinex, 2004-present)

A smug green booger who lives in your chest. Not helping you. He IS the disease, personified and taunting you with T.J. Miller’s voice.

The most recognized mascot in cold medicine. Also, most hated.

Being the most hated, however, did not stop the brand owners from turning him into a sex symbol, even with his own Tinder account.

Hamburglar (McDonald’s, 1971-present)

A character whose entire personality is stealing the product you’re selling. Originally, a creepy flasher was called “Lone Jogger” before McDonald’s made him kid-friendly.

50+ years later, still robbing restaurants in commercials with his weird but iconic ‘Robble’ sound.

Instead of showing your product, show how badly somebody wants to steal it.

Stressmannetje (Belgian Public transport)

Literally called ‘little stress guy’ is a small, annoying doll figure that makes your commute hell. Traffic, delays, chaos. That’s his whole deal. And you guessed it: a video game where you can play as the little stress guy.

Belgian transit’s answer: avoid the stress guy, take public transport.

Aleksandr the Meerkat (Compare the Market, 2009-present)

Built a billion-pound insurance brand on a dad joke about people confusing “market” with “meerkat.”

Aleksandr is posh, Russian, slightly annoyed, runs a fake “Compare the Meerkat” website that has nothing to do with insurance. Now has a whole family, including Sergei the Wombat.

15+ years running. 6.5 million meerkat toys given away. Books published. Movies made. All from a language pun.

Other baddies include Coolspot, Duo, (duh), Charlie Tuna (Sorry Charlie) and well … many more.

Why this works

Here’s what most brands miss: conflict is more memorable than comfort.

The villains do three things most “nice” mascots don’t:

1. They make the brand enemy visible. Stress isn’t abstract when it’s a little toy screaming it you. Mucus isn’t vague when it’s a smug blob.

2. They are highly entertaining. Will the Noid ruin this pizza? Will Duo catch you? Will the Hamburglar get away with it? Helpful mascots just... exist. Villains create tension.

3. They are shareable. “Friendly helpful mascot” gets zero screenshots. “Psychotic owl threatening conjugation murder” goes viral instantly.

The option most brands ignore

You don’t need a villain mascot.

Tony the Tiger, the Michelin Man, the Geico Gecko—all helpful, all successful.

But if you’re competing in a crowded category where everyone looks the same, friction might be your fastest path to memorable.

Most brands optimize for:

• Friendly ✓

• Helpful ✓

• Professional ✓

• Inoffensive ✓

Translation: forgettable.

The mascots people actually remember are causing mayhem, stealing products, or threatening users. They’re the opposite of what you’re supposed to want.

And in the age of infinite AI-generated content, being opposite might matter more than being perfect.

That’s what I did with Marcel. I knew that I couldn’t make the perfect mascot. I knew AI is still messy. So I leaned in on it. I made him goofy, weird, sometimes a bit crazy.

Subscribe now

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

If you have a mascot, you’re already 6-15x more memorable than brands relying on logos alone. Research from Ipsos backs this up - characters trigger memory structures that abstract symbols or typography simply can’t match.

So congratulations, you’re ahead if you’re even considering a mascot.

But here’s the question: how memorable is your mascot actually? And what if you’re considering designing one? How can you test before you invest?

We’re bombarded daily with amazing examples from characters like Duo and Ronald McDonald but how do we build these characters?

There’s a massive difference between “we have a character” and “we have a character that people can’t forget.” The gap between a generic dog render and the Geico gecko isn’t just creative execution - it’s memorability mechanics.

I was already using a bunch of ‘stress-tests’ in my mascot studio, but I decided to build (read: vibecode) a custom tool so you can test yours, for free. It’s based on 3 simple but crucial tests:

• Blur: When you’re scrolling, or driving, brands are often out of focus, this test looks at the overal color layout and clarity of your character

• Silhouette: Truly iconic characters have a strong shape, even when it’s dark.

• Fragment: You don’t always see the full character, sometimes it’s just walking away, sometimes you’re driving by. Fragment looks at small parts of your mascot in isolation to see if it still holds up.

Most brands never find out until they’ve spent six figures on activation. I built a free testing app so you can know before you invest. It takes a few minutes. You’ll get a score 0-10 and know where you stand.

Test it here

The honest truth: Marcel failed the test.

I ran Marcel through the BSF test myself. He got a 5.8 out of 10 😅.

Why? He’s a blue dog with accessories (hat, skull collar), but the proportions and shape still fit the “dog realm.” When blurred or in silhouette, he doesn’t scream “THAT’S MARCEL” - he screams “that’s... a blue dog?”

The 0-10 spectrum:

0 points: A generic render of a dog character. It could be anyone’s dog. Stock footage vibes. Zero distinctiveness.

5 points: Marcel (currently). Blue dog with accessories. Distinctive color, some personality, but hard to spot from afar. No strong color blocks or iconic shapes.

10 points: Totally unique character with impossible-to-confuse proportions, distinctive shape, and clear color blocking. Think early Mickey Mouse silhouette or the Michelin Man’s tire segments. Instantly recognizable even as a shadow.

Let’s fix Marcel

The test revealed the gaps. So I’m iterating based on data:

• Adding stronger color blocks: black collar (added contrast) and beige mouth area to improve blur test, so it’s not just a blue furry beast.

• Adding more distinct features (white gloves as a nod to the OG character Mickey mouse) and a more unique shape to improve silhouette and fragment score

• Custom typeface on the hat that becomes part of his identity and improve fragment score

The result? 7.5! He’s now ranking among the greats.

Curious where you stand? Upload your character. Get your BSF score. See how you rank. It’s free and takes 3 minutes.

Show me!

Here’s the thing: 6+ is good enough

You don’t need a perfect 10 to build a great mascot. The Aflac duck, Geico gecko, and Aleksandr the meerkat aren’t silhouette superstars. They’re recognizable animals (okay, maybe not the Aflac duck) that look... like animals.

But they became iconic through:

• Unique voice (Aflac’s quack, the gecko’s Cockney accent, Aleksandr’s Russian accent)

• Quality creative (actually entertaining ads, not just brand messages with a character slapped on)

• Consistent above-the-line spend (TV, billboards, sustained visibility over years)

• Years of commitment (not launching and abandoning after 6 months when the CMO changes)

The BSF test tells you if you’re starting from 0 or 5 or 8. But it doesn’t tell you if you’ll win. The memorability mechanics get you in the game. The voice, creative execution, and media strategy determine if you dominate it.

A 6+ score means you have a solid foundation. Below 6? You should probably redesign before you pour media budget into activation. Between 6-10? Optimize the other layers.

Layer 2: Sonic distinctiveness

Visual shape gets you noticed. Voice & sound get you remembered.

Ask yourself:

• Does your character have a unique voice (not “friendly corporate narrator #47”)?

• Can people identify your brand from 3 seconds of that voice alone?

• Is the voice consistent across every touchpoint (not different voice actors in different markets)?

Voice is a memory structure as powerful as visual shape. Maybe more powerful as the Aflac duck and Meerkats seem to prove.

The fame/uniqueness paradox

Jenni Romaniuk’s fame/uniqueness framework is gold for measuring existing distinctive assets. You survey your market: “How many people know this asset?” (fame) and “How many brands does it make you think of?” (uniqueness).

High fame + high uniqueness = powerful, distinctive asset. Low fame + high uniqueness = weird but unknown. High fame + low uniqueness = generic.

But this framework is terrible for developing new mascots. Why? Because you have no fame yet. People won’t recognize the asset so you’ll score low on all of these attributes.

The better approach when developing a brand asset over time

1. Test basic visual mechanics (BSF) - aim for 6+ before you activate

2. Test sonic distinctiveness - can the voice carry recognition alone?

3. Launch and iterate - does it make people feel something? (funny, cute, weird, annoying in a good way?) Are we getting traction with launch posts? Follow the comments to build a strong asset.

4. THEN build fame through consistent activation, high-quality creative, and media spend

5. THEN measure with fame/uniqueness framework 1-2 years after launch

That’s it. A small glimpse into what it takes to build a distinctive character. What do you think of Marcel 2.0? And of course, feel free to share the results of your test in the comments. I’m curious to see how your mascot ranked and if you have any questions.

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

Subscribe now

Hey brand-builder!

Here’s a stat that should change how you think about branding:

Sound has 9X better odds of creating strong brand recall than your logo.

Characters have 6X better odds.

Yet only 5% of ads use distinctive brand sounds. Only 15% use characters. Meanwhile, 90%+ of ads show the logo.

Brands are massively over-investing in the weakest distinctive asset while ignoring the strongest ones.

A talking mascot combines two of the most powerful memory structures:

1. Visual character recognition (6X odds)

2. Sonic branding through consistent voice (9X odds)

Someone scrolling with sound on but not watching? Voice wins. Someone listening to a podcast ad? Voice wins. Someone half-watching while multitasking? Familiar character voice wins.

Most brands have mascots but treat them like fancy logos: static, voiceless, underutilized. The fix is simple: give them a voice. Now with Gen-AI, you can enable these characters to speak without expensive voice actors, VFX animations etc…

Design Your Mascot for Lipsync (Or Regret It Later)

Most lip-sync AI models are trained on human faces. If your mascot doesn’t have certain human-like features, lip-sync won’t work.

Design features you need:

• Clear mouth/jaw area - The AI needs to see where the mouth is

• Visible lips or mouth opening - A line for a mouth won’t cut it

• Human-like facial proportions - Eyes, nose, mouth in roughly human positions

• Defined facial structure - The AI needs reference points (cheekbones, jaw, teeth, forehead)

If you’re creating a mascot today, design it with lip-sync capability from day one. Adding it later is painful (or impossible).

The Models That Actually Work

There are three types of lip-sync models for mascots:

1. Image-to-Video (with custom audio) - Start with a still image, generate a talking video

2. Video-to-Video Motion Transfer - transfer motion to your mascot from existing footage (or film yourself)

3. Lip Syncing/Dubbing - Add voice to existing mascot videos

TYPE 1: IMAGE-TO-VIDEO MODELS (with Custom Audio)

VEED FABRIC - The Reliable Workhorse

What it is: Image + custom audio = talking mascot

Best for: Static shots where your mascot is talking without any big movements or complex scenes

How it works:

• Upload your mascot image (neutral pose, mouth clearly visible)

• Upload your custom audio (this can be AI-generated or recorded)

• Get a talking mascot in ~60 seconds

The good:

• Most reliable for clean static shots

• Natural mouth movement and gesturing based on audio sample

• Fast generation

• Simple workflow

The limitations:

• No control over camera movement

• Backgrounds usually stay static

Note on Hedra: Another image-to-video option exists (Hedra), but it’s significantly less reliable with non-human characters. Fabric wins for mascot work.

KLING AVATAR - For more complex movement

What it is: Image + custom audio + prompt guidance = controlled mascot movement

Best for: Any shot requiring specific movement, gestures or background.

How it works:

• Upload your mascot image (mouth clearly visible)

• Add your custom audio

• Write prompts to control motion (”slight head tilt,” “dancing,” “walking toward camera,” “excited gestures”)

• Get movement + lipsync together

The game-changer: This is the only image-to-video model that lets you control the motion and shot composition through prompts.

The good:

• Full prompt control over movement (subtle or complex)

• Can handle everything from gentle nods to full dancing

• Dynamic camera work is possible

• Natural integration of movement + lipsync

The limitation:

• Requires character reference pack setup

• More variables to control than Fabric

• More complex workflow

TYPE 2: VIDEO-TO-VIDEO MOTION TRANSFER

MODELS: WAN ANIMATE / RUNWAY ACT 2 / KLING MOTION TRANSFER

What it is: Your filmed performance + mascot image = mascot doing your exact movements

Best for: Complex human gestures and detailed motion when you need specific movements

How it works:

• Film yourself performing the exact motion/gesture you want (or use existing footage from movies), including the audio

• Match the first frame of your mascot with the first frame of the reference video (so it can really map onto the character movement) and upload it as an image reference

• Model transfers your motion to your mascot + adds lipsync

The good (when it works):

• Most natural human-like movement possible

• Perfect for specific complex gestures you can’t easily prompt

• Captures nuance and timing from real performance

• Great for detailed hand gestures, body language, and complex actions like dancing

The reality:

• Very buggy across all providers

• Hit or miss results

• Can produce amazing outputs or complete failures

• Inconsistent quality across generations

• Requires multiple attempts to get usable results

Example: The “please just keep the tail” praying hands gesture - easier to film than prompt, but expect to generate 5-10 times before getting a keeper + some work to get the matching frames

Input video

Wan animate result (+ Kling O1 edit for background)

Models in this category:

• Wan Animate

• Runway Act 2

• Kling Motion Transfer

• Others emerging regularly (Luma Modify)

Marcel says: “Motion transfer is like playing the lottery. When you win, it’s gold. When you lose, it’s a three-legged nightmare. Budget extra time and patience.”

Overall, this category of models is very promising, as human driven acting still gives the most nuanced results, but currently it seems to do better at transferring full body movement (like dancing) as shown in this recent amazing example from Ink.

TYPE 3: LIP SYNCING / DUBBING MODELS

SYNC / HEYGEN - The Post-Production Fix

What it is: Existing video + new audio = re-synced lipsync layer

Best for: Adding voice to finished videos, translations, and audio replacements

How it works:

• Take any existing mascot video (even without audio originally)

• Upload new audio file

• Tool layers lip-sync on top to match the new audio

The good:

• Fast audio addition to existing content

• Works with any existing video

• Great for translations and voice changes

• Good for fixing audio issues in post

• You can get a more complex scene with top-tier video models like VEO 3.1, Kling 2.6, and then overlay the talking

The limitation:

• Lipsync can feel “layered over” the original

• Less seamless integration than image-to-video or motion transfer

• You can see the seams if you look closely

• Doesn’t add motion in the scene, only talking movement (you can specify if you want lip, face, or head movement)

Tools in this category:

• Sync

• Heygen

• Similar dubbing services

Marcel says: “Gets the job done when you need it. Not as clean as building voice in from the start, but sometimes you just need to fix it in post. That’s life.”

Why Most Video Models Don’t Work for Mascots

Models like VEO 3.1 (4 coming soon), Kling 2.6, and now Seedance Pro 1.5 can generate impressive videos with audio.

For mascots? It doesn’t work (for now at least).

These tools generate videos with random AI voices. Different voices every generation. Generic. Uncontrollable. + The added bonus of the typical AI compressed voice that screams ‘AI’ to everyone.

For mascots, this breaks everything:

Your mascot needs its voice. The same voice. Every single time.

That voice might be:

• AI-generated character voice (ElevenLabs, etc.)

• Voice actor recording

• Your own voice

• Doesn’t matter, but it must be consistent

Brand recognition = voice recognition.

If someone hears your mascot and doesn’t immediately know it’s your brand, the voice is wrong.

The Three-Type Approach

This is why we use the three model types covered above. They all accept custom audio input:

• Type 1 - Image-to-Video: Fabric, Kling Avatar (Image + your audio)
  Type 2 - Motion Transfer: Wan Animate, Act 2, Kling Motion (Video + your audio)
  Type 3 - Dubbing: Sync, Heygen (Video + your new audio)

These tools prioritize custom audio because that’s what builds brand memory.

Generic AI voices = no distinctive asset. No brand equity. No memory structure.

Marcel says: “See those fancy video models everyone’s hyping? They give me a different voice every time. That’s not a mascot. That’s digital schizophrenia. Hard pass.”

Subscribe now

What 2026 Looks Like

This year, we went from “can it move its mouth?” to “can it gesture naturally.”

Next year? I expect 2 things:

• Custom audio training for top-tier video models like VEO 4 (perhaps via Elevenlabs)

• Real-time avatars with custom voice

The tech is already emerging:

• Live mascots in video calls

• Interactive customer service avatars

• Real-time streaming characters

• Mascots responding to live audience input

The brands that have their character voice systems built now will dominate.

The brands still treating mascots as “just logo variations” will scramble to catch up while everyone else is already live and interactive.

Voice isn’t optional anymore. It’s infrastructure.

For a broader view on mascots in 2026, check out this article

That’s it! Thanks for reading. And happy 2026!

Subscribe now

Want to build your brand universe?

I help brands create character-driven engines powered by AI.

Book a 20-minute call | See examples

Subscribe to Marcel’s Lab - The only newsletter on building character-driven brands powered by AI. When content is infinite, characters are your moat.

A year ago, generating a consistent mascot with gen-AI meant:

• Painstakingly spinning the slot machine to get multiple angles from a single image starting point

• Training multiple generations of LoRAs from a crappy dataset of images to get half-baked consistency

• Still getting 47 variations of “close but wrong" when taking that model for a spin

• Abandoning your creative vision for whatever worked

Today? We’re in a completely different game.

Let me show you with Marcel himself.

That’s the gap we closed in 12 months.

• Consistent character

• Extremely detailed

• World understanding (e.g. added the flashlight)

• Simple instructions (no crazy prompting architecture)

• 4K quality (just zoom to see the difference)

Not to mention the ability to render text consistently, combine multiple images, upscale, etc..

The Creativity Explosion

Sketch → Full Universe

Ink’s sketch-to-world example shows what’s now possible: a napkin doodle to complete a character world in one session.

Node-based workflows finally match how brains actually work. No more fighting with text prompts. Show it. Build it. Move on.

Models like Nano Banana Pro (or Gemini 3 Pro) understand the intent and context. The shift to multimodal input (image(s) + basic instructions) removed the weight of “prompt engineering.” You show it what you want. It gets it.

You can now combine reference styles, locations & characters and blend them like an alchemist. The model does the heavy lifting.

Want to give it a try? Head over to Marcel’s sandbox and try it for free.

Consistency: Actually Solved

This was THE blocker. Editing models like Nano Banana Pro and Seedream cracked it.

Same character. 50 different poses. Zero drift. No more custom-trained models. You can generate a single character sheet from one image and use that character sheet anywhere to generate consistent characters.

Video Control Got Surgical

First frame/last frame control changed everything. But hybrid performances and advanced video editing pushed it further.

We went from slot machine to smart planning:

• Storyboarding tools plan entire narratives

• Direct scenes like an actual filmmaker with first/last frame and video references

• Control character movement frame-by-frame

• Make your character speak, driven by voice samples (can be human-recorded or AI-generated) with models like Hedra, Kling Avatar or Fabric

See the full behind-the-scenes breakdown of my latest short film, ‘The Bubble’ to see this in action in the article below.

Snippet from my latest work for AI sales assistant Donna. Full video here.

What is still broken?

• Audio quality on img to video models like Wan 2.6, Kling 2.6, and VEO 3.1 still sounds robotic. Currently, it’s better to use Lip Sync models as listed above.

• Character consistency in animation, especially facial features, is still an issue in video generations. Models often default to generic cartoon eyes and mouths.

Coming soon: As VEO 4 will have more understanding, like Nano Banana pro, I’m assuming it will be a lot better at solving both problems (audio and character consistency beyond the understanding of the img reference).

👀 Looking ahead to 2026

Here are some things I’m looking forward to in 2026 that will truly level up your mascot game.

World Building

This is not yet there yet, but I think it will be big in 2026. World models like Genie 3D (Google Deepmind) and World labs (Gaussian splatting) let you create navigable 3D spaces from 2D characters.

Your mascot will be able to exist in a consistent place, not just a flat image.

Character universes, not just character images.

Live Avatars

Real-time avatar solutions are showing serious signals. Genie, Avatar models, Runway experiments.

Zero-latency character performance isn’t solved (and mostly awkward, trust me). But we’re 6-12 months away, not years.

So What Does 2026 Look Like?

If you were subscribed to this newsletter, you might recognize the playbook for character-driven brands in the era of AI:

• 🐨 They will become more Distinctive

  • More unique, creative characters and worlds. The barrier to distinctive IP just collapsed.

• 🎪 They will become more Entertaining.

  • Unique stories driven by unique characters and IP. No infinite slop factory.

  • Longer form, higher quality, actually watchable content.

• 🕹️ They will become more Playable

  • Live avatar solutions: Real-time character interaction moves from experimental to standard.

  • Characters will cross more and more into physical space. Robotics is finally getting interesting. Not because the tech got perfect, but because designers stopped trying to make it perfect. Brands like Sunday and Disney are getting it right.

And of course, more slop (Merriam-Webster’s word of the year. We called it.)

Marcel: Physical me would absolutely have weird quirks. That’s not a bug, that’s branding.

Here’s the thing:

When everyone can make infinite content, creativity becomes the ONLY moat.

2025 proved that AI-generated characters can be:

• Consistent and qualitative enough for brand work

• Creative enough to stand out

• Efficient enough to scale

• blended into physical reality

The brands investing in distinctive character IP and universes right now will own 2026.

Everyone else will be making slop.

Marcel: And I’ll still be here, being handsome and correct about everything.

Ready to build your brand universe?

I help companies create distinctive brand universes and characters using Gen-AI. The kind that is actually used (not stuck in a Figma file somewhere). If you’re exploring this for your brand, book a 20-min call or check out some examples here.

And of course, you can subscribe to the newsletter to learn how to build a strong brand powered by Gen-AI.

Subscribe now