I start every PM coaching session with the same provocation:

“If you think your job is to gather requirements and create specifications for engineers, then do not call yourself a product manager.”

The room gets quiet. A few people shift in their seats. Most of them have been doing exactly that — and doing it well. They have clean PRDs, well-groomed backlogs, tight sprint plans, and stakeholders who like them.

And yet, they are not product managers. Not really. Not yet.

Let me explain with a bakery.

The Bakery Analogy

Imagine you walk into a bakery and meet two people.

The first person wakes up at 4am every day. They know every recipe. They bake sourdough loaves with perfect crust. They measure flour to the gram. They never miss a step. Their croissants are flawless.

The second person also knows how to bake. But that is not what they spend most of their time on. They are thinking about which neighborhoods to open a second location in. They are asking why the muffin shelf clears out by 8am but the rye bread is still there at noon. They are figuring out whether to add a lunch menu or a subscription box. They are talking to regulars to understand what would make them spend more.

The first person is a baker. A great one.

The second person is a bakery owner.

Now here is the question I ask my students: Which one are you?

Most PMs at large companies are bakers. Excellent bakers. Technically skilled, process-fluent, detail-oriented. But they are baking what someone else decided should be on the menu — and they rarely ask whether it is selling.

What the Bakery Owner Does That the Baker Does Not

The bakery owner does not stop baking. The product still has to ship. The PRD still needs to be written. The sprint still has to be planned. That work does not go away.

But the owner starts with a completely different question.

Not “what do the engineers need from me today?”

But “what will make this business grow?”

Every decision flows from that. The recipes they invest in, the items they put on the menu, the things they discontinue — all of it is in service of a revenue outcome.

This is exactly how a product manager at a large enterprise company needs to think.

Your roadmap is the menu. Your sprint plan is the kitchen schedule. Your PRD is the recipe card. These are tools in service of a goal — not the goal itself.

The goal is revenue growth.

Baking Revenue Into Your Product Plan

Here is where most early PMs get stuck. They accept the revenue goal as a given — “yes, of course we want to grow” — but they never actually connect their daily decisions to that outcome. Revenue lives in the finance deck. Product lives in Jira. And the two rarely meet.

That is a mistake. And it is fixable.

There are six concrete ways a PM can influence revenue. Think of them as six different bakery growth strategies.

1. Expand Into Your Ideal Customer Profile — Fill More Tables

Your bakery has regulars. The growth motion here is to find more customers exactly like them — same neighborhood, same habits, same tastes.

As a PM, this means making the product so good for your core customer that sales has a repeatable story to tell. You identify the friction points that slow down new customer onboarding. You build features that strengthen your value proposition against competitors. You make it easier for a new customer to become a happy customer, fast.

This is the most common growth lever at large B2B companies, and it is the one PMs most directly influence.

2. Retain Customers — Stop the Leaky Oven

Every bakery loses customers. A few move away. Some find a new favorite. Some just stop coming in. The owner obsesses over why.

Retention is the lifeblood of SaaS. Acquiring a new customer costs far more than keeping an existing one — and for a $100M ARR business, every 1% reduction in churn saves $1M in revenue that would otherwise walk out the door.

Your job as a PM is to understand exactly where and why customers churn. Is a specific feature broken? Is a certain segment churning faster than others? Is the onboarding experience failing new users before they ever reach value? Find it. Fix it. And build every new feature with stickiness in mind.

Faster time-to-value is not a UX nice-to-have. It is a retention strategy.

3. Sell More to Existing Customers — Upsell the Regulars

Your regulars come in every morning for a coffee and a croissant. Have you tried offering them a loyalty subscription? A weekend brunch box? A second coffee at half price?

In product terms: cross-sell and upsell.

Cross-sell means offering adjacent products to the same customer — adding a marketing automation module to a CRM customer, for example. Upsell means extracting more value from the existing relationship — moving a customer from 100 seats to 1,000, or from a Silver plan to Gold.

There is also the free-to-paid conversion motion: moving users who signed up on a free tier to a paid plan. This requires understanding exactly what value will compel them to pay — and making sure that value is unmistakably clear and just out of reach on the free tier.

All of this is a PM’s job. Not sales’s job alone. Yours.

4. Create a New Customer Segment — Open a Catering Line

Your walk-in retail customers are loyal. But one day someone asks if you do corporate catering. You had not thought about it. Same kitchen, same bakers, same recipes — but a completely different customer, occasion, and revenue model.

This is the new segment play. You take your existing product and ask: who else has this pain? An employment verification product for HR teams might work just as well for property managers verifying tenants. A fraud detection tool built for banks might be adaptable for insurance companies.

As a PM, your job is to evaluate which adjacent markets your product can serve, how much the product would need to change, and whether the revenue opportunity justifies the investment.

5. Create New Products — Add a Lunch Menu

The bakery is doing well at breakfast. But there is a gap — customers come in at 8am and again at noon, but you are not capturing that second visit. So you build a lunch offering.

For mature B2B products, new product lines are how companies unlock the next phase of growth. Think Docusign expanding into contract lifecycle management. Think Adobe adding marketing automation. The existing customer base trusts you; the question is whether you can solve another problem they have and get them to open a different budget line for you.

This is harder, and it requires a different kind of PM work — validating a new value proposition, understanding a new persona, and building something net new. But it is a legitimate growth lever and one that large companies pursue regularly.

6. New Product in a New Market — Open a Second Bakery in a Different City

This is the hardest move. New product, new customers, new market, new competition. No brand recognition. No shortcuts.

At large companies, this often happens through acquisition rather than organic build. It is worth understanding as a strategic concept — but for most early PMs, this is not where you should be spending your energy.

So Which Lever Should You Pull?

This is where strategy lives. And there is no universal answer — but there is a framework.

If your market is large and you are not yet the market leader, the priority is usually to expand into more ICP customerswhile aggressively reducing churn. These two together compound into meaningful growth.

If your market is mature or highly competitive, look at cross-sell and upsell. Your existing customers trust you. Selling to them is easier and cheaper than acquiring new ones.

If you are hitting a ceiling in your current market, look at adjacent segments or new products. This requires more investment, more research, and more patience — but it is how established products sustain growth curves.

The critical skill is knowing which one is right for your business right now — and letting that choice drive your roadmap, not the other way around.

This is the thinking I go deep on in my guide: B2B Product Manager’s Guide to Revenue Growth. It covers all six levers with examples and walks through how to decide which path to take given your market position.

The Mindset Shift

Let me bring this back to the bakery.

The baker’s job is to execute. Flawlessly, consistently, every day. That job matters.

But the bakery owner’s job is to decide what gets baked — and why — based on what the business needs to grow. They use the baker’s skill in service of a commercial outcome.

As a PM at a large company, you are surrounded by execution machinery. Engineers, designers, QA, program managers — all of it pointed at shipping software. You are not the only one who can operate that machinery. You are the one who should be deciding what it produces.

That means you need to know your market. Your customers. Your competitors. The revenue levers available to you. The growth strategy your company is pursuing.

Your roadmap, your PRD, your sprint plan — these are the recipe cards. They are not the strategy. Strategy is deciding which items belong on the menu and why.

Stop being the baker.

Start thinking like the owner.

Deepak Deolalikar coaches B2B product managers at enterprise companies to build products that generate repeatable revenue with high retention. Read more at businessofproducts.com.

Share

Someone answered in about four seconds. Confident. Clean. Everyone nodded and the meeting moved on.

I was still thinking. I had a question about whether that deal was actually at risk or just loud. I never asked it. By the time the question was shaped the way I wanted it, we were three topics later.

I drove home that evening feeling like something was wrong with me.

If you’re a thinker, you know that drive home.

You’ve probably also collected the feedback that comes with it. Be more decisive. Trust your gut. You’re in your head too much. Just say the thing.

So you try to become the four-second person. I tried for years.

It doesn’t work. And here’s the part nobody tells you — it makes you worse. You start pushing out opinions you haven’t finished forming, in a voice that isn’t yours, and now you’re neither fast nor right.

Then I realized I had the diagnosis backwards.

The thinking was never the problem

Look at what our product job actually is.

Customers say one thing and do another.

Sales pushes a feature that saves one deal and wrecks the roadmap.

Engineering surfaces a constraint that changes the whole shape of the solution.

Leadership drops a new priority in during week three.

None of that rewards a fast reflex. It rewards someone who holds five contradictory inputs in their head and notices which one doesn’t fit.

That’s you. That’s the whole thing you do naturally.

The questions that come easily to thinkers -

• What are we assuming here

• What happens if we’re wrong,

• Who actually asked for this

• Are the questions that stop a team from burning a quarter on the wrong build.

  
  Most rooms have plenty of people willing to answer quickly. Very few have someone willing to slow the room down.

You are not a slow version of the confident person. You’re a different instrument.

So what actually needs fixing

One thing. Not your personality. Your stopping rule.

Deep thinking stops helping the moment it stops changing the answer. That’s the whole line. Everything before it is judgment. Everything after it is just delay wearing a lab coat.

Think about checking the weather before a trip. Forecast says sixty percent rain. You check again the next day. Still sixty. You were packing the jacket either way. Refreshing the app doesn’t change what goes in the bag.

Most product decisions hit that point way earlier than we admit.

So the only question I ask now, and the one I ask every PM I coach:

What would I need to learn for my answer to change?

If nothing comes to mind, you’re done. Not because you’re certain. Because more thinking has stopped paying.

Then say it out loud before it feels finished. This was the hard one for me. I used to wait until my thinking was airtight, because I wanted to walk in with the answer. But nobody’s waiting for your finished answer. They need your half-formed one early enough to shape it with you. “Here’s where I’m leaning, here’s what’s still bothering me” moves a team further than a perfect memo two weeks later.

That’s it. That’s the whole adjustment. You keep the thinking. You just stop letting it run without a clock.

What I’d tell my younger self on that drive home

Nothing is wrong with you.

The room did need a fast answer that day. It also needed the question you didn’t ask. Those aren’t in competition. You just have to actually put yours on the table before the topic changes.

You don’t have to become the four-second person.

You have to become the person who knows when they’re done.

Share

It wasn’t the content. It was the delivery. Flat pitch, rushed pace, voice trailing off right at the point that mattered most.

I used to think good ideas sell themselves. They don’t. How you say something carries almost as much weight as what you say — whether you’re a founder pitching investors, a PM presenting a roadmap, or anyone trying to get a customer to care.

And here’s the good part: speaking well isn’t a personality trait. It’s a skill. You can learn it, the same way you learn any other part of the craft.

Came across this TED Talk by Julian Treasure - “How to Speak So That People Want to Listen.”

He breaks down the habits that make people tune out (he calls them the seven deadly sins of speaking — gossip, judging, complaining, and a few others we all fall into without noticing), then gives you a simple foundation to speak from: honesty, authenticity, integrity, love. After that, he gets practical — how pitch, pace, and volume change whether a room leans in or checks out.

Watch it before your next pitch, not after.

Your best idea is only as good as the room’s willingness to listen to it. Worth the 18 minutes.

Here is the TED talk

Have a great long weekend in the USA. Happy July 4th.

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Background

At iCharts, I was employee #2. The first thing we had to figure out wasn’t the product. It was who we’d sell it to first.

I was a student of Steve Blank at Berkeley. So I already believed in selling to a niche. The hard part was figuring out which one.

Our software was a data visualization tool. Upload a CSV, or connect to a data source, and create interactive HTML5 charts you could embed on your site — public or private. It was horizontal. Anyone could use it.

That’s the first mistake almost every founder makes. We believe our product is so good that anyone can use it.

Sure. But you can’t sell to everyone.

Fortunately, I’d already learned that lesson. So I talked to 80 prospective users across a wide range of domains. Associations like AMA, who put out charts and data regularly. News organizations like Wall Street Journal and Forbes, who wanted interactive visualizations for their digital properties. Analyst relations teams at public companies, sharing results with investors. Market research firms, sharing survey results with clients. Business analysts inside enterprises, sitting on Excel sheets, trying to make charts for presentations.

(Yes, Excel and PowerPoint make charts. But they’re not interactive, they’re a pain to embed on a website, and they don’t update in real time when the data changes.)

After all those conversations, I grouped them into segments — associations, enterprises, market research, media, government agencies. Then I plotted each one on a 2x2: extent of pain, and willingness to pay.

Market research won. They had a real, painful problem — visualizing survey results in real time — and they were ready to pay to get rid of the manual effort and give their clients faster answers. That narrow focus is what got us our early revenue.

Why going narrow gets you to revenue faster?

Here’s the actual goal for a startup: repeatable sales. That means landing your next customer takes less effort than landing the last one. Not more. If you can’t get there, you don’t have a business. You have a string of one-off deals, and each one means reinventing the pitch from scratch.

Niching down is how you get repeatability. Sell to ten different types of customer and you need ten different pitches, ten different messages, sometimes ten different channels — because a hospital ops director and a real estate broker don’t read the same publications or care about the same proof points. That’s not a go-to-market plan. That’s ten of them, running at once. You can’t afford one yet, let alone ten.

Pick one niche, and your message stays the same from the first call to the fiftieth. That consistency is what actually lowers your cost of acquiring a customer. Not a clever ad. Not a bigger budget.

It changes how the customer experiences you, too. If you build a generic CRM, then each customer type or segment has to do the work of figuring out if it’s even for them. But, if you say we are “a CRM for real estate agents,” then your chosen segment of real estate agents will understand that this is built specific to them. Friction gone before the sales call even starts. That’s the difference between a long sales cycle and a short one.

A niche isn’t a limitation. It’s a filter.

If you’re a new B2B founder with an idea, the most important decision you’ll make early on has nothing to do with the product. It’s deciding which narrow segment of people you sell to first.

That segment isn’t an industry. It isn’t a company size. It’s four things, all true at once:

- They have a huge problem they need to solve today

- Their current alternative is bad

- Something has them motivated to act now

- They have the budget to actually pay you

A niche, for an early-stage founder, is just the set of customers sitting at that exact intersection.

Most founders skip this and go looking for a market instead. Niching down feels like settling — like you’re limiting your ambition before you’ve even started. To complicate matter, their investors are asking for a big TAM. Don’t confuse the two. TAM is the end vision. Niche is where you get momentum.

Founders who get to revenue fastest are the ones who find that narrow intersection first, then expand. The ones who try to serve everyone from day one spend months chasing interest that never converts.

A niche answers one question: who, specifically, sits at that intersection? Not “enterprise companies.” Not “small businesses.” A specific job title, at a specific kind of company, dealing with a specific situation. The more specific you get, the easier it is to find them, talk to them, and sell to them.

Here’s what each of those four conditions actually looks like.

1. They have a huge problem they need to solve today

Not someday. Not “on our radar for next year.” Today. If a prospect can shrug and say “we’ll get to it eventually,” that’s not your niche. That’s politeness, and politeness doesn’t convert.

I teach founders to think in three layers: vitamin, painkiller, chemotherapy.

A vitamin is nice to have. It gets cut the second budgets tighten, because nobody’s business breaks without it. A painkiller fixes real, current friction, and gets renewed first, because the pain doesn’t go away on its own. Chemotherapy is different. The customer isn’t shopping around. Isn’t comparing features. Isn’t waiting for next quarter’s budget cycle. They need it today, because the alternative is existential — a compliance deadline, a security breach, a vendor shutting down, a regulatory mandate with a hard date. They don’t ask “is this worth it?” They ask “how fast can you start?”

As a new founder, you can’t afford to sell vitamins. You usually can’t win on painkillers either, not yet. There is also a lot of competition in pain killers. You don’t have the brand or the case studies to compete on “this would help.” What you can win is chemotherapy — the customer who needs it fixed today, who isn’t negotiating on price, because doing nothing costs more than anything you’d charge. And they do not have the time to shop around.

Here’s what that looks like. Say you’re selling a tool that flags expired vendor certifications before an audit. Most companies will say “good to know, we’ll look into it next quarter.” That’s a vitamin. But the company that just got a letter saying their audit moved up two months, and three certifications are already expired? They’re not comparing you to competitors. They’re asking how fast you can start. Same product. Completely different customer.

2. Their current alternative is the villain

Every customer is already solving this problem somehow. A spreadsheet held together with duct tape. A manual workaround three people maintain by hand. A legacy tool everyone complains about but nobody’s replaced.

That status quo is your villain. And here’s the gift: you don’t have to convince them it’s bad. They already know. They live with it.

Your job isn’t to invent the pain. It’s to find the niche that’s already furious about it.

Check social platform like Reddit, X or Quora. Check our G2 or Capterra reviews. You will find users complain about their current solution.

3. They’re motivated to act now

A bad alternative on its own isn’t enough. People tolerate bad tools for years. You need a trigger — something that changed recently and took “we’ll deal with it later” off the table.

A new compliance deadline. A new VP who won’t accept the old way. A competitor that just pulled ahead. A headcount cut that makes the manual workaround impossible to staff. Something concrete. Something dated.

If you can’t point to a specific reason this niche needs to move now, you’re early to their problem. Not wrong — early. And early is expensive when you’ve got no runway to wait.

4. They’re willing to pay

This is the one founders skip, because it’s the one that tells the truth. Interest is cheap. A demo request is cheap. “This looks promising, keep us posted” is cheap.

Payment is the only signal that counts. Budget they already control, today, that they’re willing to commit before your product is even fully built. If a niche has the pain, the villain, and the urgency — but no budget, or no authority to spend it — you’ve found believers, not customers.

What to do with this?

A niche that checks all four boxes is small on purpose. A small group of people in real pain is easier to find, faster to reach, quicker to close than a broad market of people who are mildly curious. You can name them. Call them. Learn their language well enough that your copy makes them feel understood, not marketed at.

Close even a handful of customers inside that niche, and you have something worth more than a clever idea: proof. Proof that people in real pain will pay to solve it. That’s what funds your next round, earns your next referral, and tells you exactly which adjacent niche to chase next.

Don’t write down generic segments like “mid-market companies.” or “Manufacturing companies”. Be as specific as possible.

Here are some examples :

1. Payroll compliance for healthcare staffing agencies

• HR ops manager at a mid-sized healthcare staffing agency

• A new state wage-transparency law just took effect, and their manual payroll process can’t track it

• Their spreadsheet-based system was fine until the rule changed overnight

2. Inventory reconciliation for specialty e-commerce brands

• Operations lead at a DTC brand selling through 3+ sales channels (Shopify, Amazon, retail)

• Just had a stockout during a viral spike, and manual reconciliation across channels couldn’t keep up

• Now leadership is asking why inventory wasn’t caught in time

3. Contract renewal tracking for boutique law firms

• Office manager or junior partner at a 10-30 person firm

• A client churned after a renewal date was missed, and it became a visible, embarrassing miss

• The firm has no system beyond someone’s calendar reminders
  

Write down the actual job title, the actual trigger event, the actual budget line your first customer would pull from. Then go find five to ten people who match all four conditions — not five hundred who match one.

Ask your early customers, why they bought from you. That will help you define your niche.

Start narrow. Get revenue. Expand from there.

New: Explain intent, the tool does it.


Now, some industries have caught up really fast. Software for example. It’s just writes the code for you.

Sales and marketing is next. But what are some other domains where the new model could be leveraged.

This article from Julien at Sequoia give a good clue.

Find your next opportunity.

Here is this week’s Gem.

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Same blank stares.

Same response: “But that’s what everyone is asking me to do.”

The New Jira Ticket Is “Add AI to This”

A year ago, PMs were drowning in backlog grooming and JIRA engineering. Today they’re drowning in AI feature requests. Different packaging. Same problem.

Every exec wants an AI roadmap. Every competitor is announcing something. And PMs are scrambling to bolt on chatbots, copilots, and “intelligent” features without asking the basic question: does this make money?

That’s not a product strategy. That’s a feature factory with a better GPU.|

Now, there is no harm in learning about AI, the models, how they work, limitations. In fact, that is a necessity for all PMs. Not just for productivity gains, but to see the potential of real impact to their customers.

But do not think that just knowing AI will make you a stellar PM.

Your job hasn’t changed. Build a business using your product. AI is a tool, not a strategy. And to do that well, here are the ten skills that matter more than ever.

1. Experience Design — AI Makes This Harder, Not Easier

Everyone assumes AI will make products smarter and easier to use. Often the opposite happens.

AI outputs are unpredictable. They fail in ways users don’t expect. A chatbot that gives a wrong answer with total confidence destroys trust faster than a feature that just doesn’t work.

Good UX in the AI era means designing for failure gracefully, setting honest expectations, and making sure users always know what to do next when the AI gets it wrong. If you can’t articulate the user’s path when the AI fails, you don’t have a product yet.

Pick any AI product you use today. Notice how frustrating it is when it confidently gives you the wrong answer. That’s a design problem. Learn to solve it.

2. Business Models - Know What the AI Actually Costs You

LLMs are expensive to run. Inference costs are real. And your customers’ willingness to pay hasn’t magically gone up because you added AI.

This is where most AI products fall apart. A feature that costs you $2 per user per month to deliver, buried in a $30 plan, is a losing bet at scale.

Before you build any AI feature, map the business model. What does it cost to deliver?

What does the customer get in return? Can they get 10x the value of what they pay? If yes, you have something. If not, you’re subsidizing a feature nobody truly needed.

Read Business Model Generation by Osterwalder. Then layer in the AI cost structure your engineering team is actually facing.

3. Growth Levers — AI Changes How Products Grow

The old growth playbook was: acquire users, activate them, retain them. AI adds a new wrinkle — the product gets better with use, which can be a powerful retention lever. But only if you design for it.

The PMs winning right now are identifying specific moments where AI creates an “aha” moment that a non-AI product couldn’t deliver. That’s your growth lever.

Study how tools like Cursor or Perplexity are growing. It’s not traditional acquisition. Users are converting because the product does something they couldn’t do before, not just faster. Find that moment in your product and build toward it.

4. Marketing — Buyers Are Researching Differently Now

This one is urgent and most PMs are completely ignoring it.

B2B buyers are no longer just Googling your product. They’re asking ChatGPT, Perplexity, and Claude. They’re getting AI-generated shortlists before they ever visit your website. If your product isn’t showing up in those answers, you’re invisible to a growing segment of buyers.

That’s not just an SEO problem. It’s a positioning and distribution problem.

PMs need to own the question: how does our product get found when a buyer asks an AI assistant what tools to use? That means clear, consistent, specific positioning — not vague category language. Read Obviously Awesome by April Dunford. It was written for this exact challenge, even before AI search existed.

5. Economics — AI Is Reshaping Entire Markets, Not Just Features

In some industries, AI isn’t a feature. It’s a full disruption. The question is whether it’s disrupting your competitors or you.

Entire cost structures are changing. Services businesses are being replaced by software. Software products are being replaced by AI agents. Buyer procurement cycles are shortening. Willingness to pay is shifting because the alternatives are getting better fast.

PMs who understand basic economic forces — supply, demand, substitution — will see these shifts coming. Those who don’t will be surprised when their retention numbers start dropping and can’t explain why.

Follow what’s happening in your specific industry beyond your product. The macro forces matter.

6. Statistics — AI Outputs Are Probabilistic, Your Decisions Should Be Too

This is the most underrated skill gap in product right now.

AI doesn’t give you right or wrong answers. It gives you probabilistic outputs. A model that’s right 90% of the time sounds great until you’re building a product where a 10% error rate destroys user trust or creates legal liability.

PMs who understand statistics can set honest expectations with customers, make smarter build vs. buy decisions, and evaluate whether that AI feature is actually working or just feels like it is. “Our AI improved conversion” means nothing without knowing the baseline, the sample size, and whether the result is statistically significant.

Most PMs are not trained here. It’s a real edge if you put in the work.

7. Industry — AI Is Not Hitting Every Industry the Same Way

The AI hype is general. The actual impact is specific to each industry.

In healthcare, regulation will slow AI adoption significantly. In legal, liability concerns dominate. In B2B SaaS, the shift is happening in how buyers research, evaluate, and buy — not just in how they use the product.

A PM who understands their industry’s specific constraints, procurement cycles, and competitive dynamics will make better AI product bets than one who’s just following the general AI news cycle.

Read industry-specific research. Talk to buyers about what they’re actually changing in their workflow versus what they’re experimenting with. The gap is usually large.

8. Finance — The AI Business Case Is Harder Than It Looks

Most AI features are easy to demo and hard to justify in a P&L.

Leadership will ask: what’s the ROI? What does this cost us to build and maintain? How does this affect our gross margin? If you can’t answer those questions, someone else will answer them for you — and you won’t like the outcome.

PMs need to build the business case for AI investments with real numbers, not just “this will improve user experience.” Quantify the expected retention lift. Model the infrastructure cost. Estimate the support cost reduction. Make it concrete.

I once had to model a joint product offering and recommend to leadership that we walk away because the economics didn’t work. That’s the job. Be willing to do that.

9. Pricing — AI Creates New Pricing Opportunities Most PMs Are Missing

Usage-based pricing is becoming the default for AI products because the cost structure is usage-based. But most PMs are still defaulting to flat monthly subscriptions because that’s what they know.

Think carefully about what your customer is actually getting. Is it time saved? Revenue generated? Decisions made faster? Price against the outcome, not the feature. A product that saves a customer $50,000 a year in manual work can be priced very differently than one that “enhances productivity.”

AI also creates natural upgrade triggers. A user who hits a usage limit on an AI feature is primed to upgrade in a way that a user who hits a seat limit is not. Design your pricing to capture that.

10. Sales - Customers Are Confused About AI, and That’s Your Problem to Solve

Customers are skeptical. They’ve been burned by AI demos that didn’t survive contact with real workflows. They don’t know what to trust.

Getting on calls and hearing those objections firsthand is how you build a product that actually sells. “We tried an AI tool and our team stopped using it after two weeks” is the most valuable sentence a PM can hear right now.

Sit in on ten sales calls this quarter. Listen for where AI comes up. Listen for the hesitation. Listen for what the customer is actually trying to get done versus what they’ve been sold before. Build for that gap.

The Skill That Ties All Ten Together: Judgment

Anyone can add AI to a product. The hard part is knowing when it adds real value, when the economics work, when your customer is ready for it, and when you’re just chasing noise.

The PMs who will win in this cycle combine business sense with honest thinking about what AI can and can’t do. Not the ones with the most impressive AI roadmap.

Start observing the world and the products you see everyday. Ask questions – Who is this product for? What are they willing to pay? Whats the alternate? What is the value customers receive?

A great Product Manager does much more than just writing JIRA tickets and managing backlogs. The best PMs understand business, growth, and strategy at a deep level. They think beyond features and focus on building products that create real value.

They build product sense.

Matt Schumer write this article on X which is quite eye opening. If your job involves a computer screen, then you must read this.

Here it is.

Update: Another person, Connor Boyack wrote this antidote article which is more grounded in human history and economics.

Read this too.

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Annual Product Planing

New year brings a new round of product planning. In this newsletter, lets go through how you should think about prioritizing to build your product plan and get alignment.

Based on your recent customer and market discovery, you have -

• brand new product ideas

• new feature capabilities

• new add on product ideas

• improvements to existing product

• improvements to sign up, activation or adoption

• customer requests

• internal tooling

• build robustness in your product

• automate testing

and so on.

Unless you have unlimited resources, you will need to prioritize from amongst these options.

In a @lennysan podcast with @anujrathi, he discussed the 4BBs to consider in your product planning.

• Big Bets

• Bread and Butter

• Breaking Bad

• Brilliant Basics

Let’s understand these categories first :

Big Bets (New Ideas)

These are new ideas or business models for your product.

For example :

• Introducing a self service component for your product

• Adding new or add on product

• Expanding to new industry or GEO

Big bets don’t happen that frequently. These are large initiatives in any company. Significant business impact is expected out of big bets.

Bread and Butter (Features and Improvement)

New features ideas and refinements to your existing products. This is essentially your backlog. Even within these you will have a ton of options and you will prioritize them as per the product strategy.

For example:

• UX Improvements

• New dashboards

• Improve activation flow

• Customer specific asks

• Stakeholder asks

Breaking Bad (New Directions)

This is a rare category for most but these are ideas that require redefining your business or product.

For example:

• OEM your product into another platform

• Refactor business product to a consumer version

• Offering a completely different product or going into a new market. (e.g. Uber eats)

This category of work will require significant discovery, validation and management buy in. The whole company will be involved in making this work.

Once I was involved in splitting our product into multiple SKUs, each with it’s own name, segment, price point and value proposition. This was a big effort and required every department to make changes.

Many companies had to implement Breaking Bads when Covid hit and as they had adapted to new business models.

Brilliant Basics (or Foundations)

A euphemistic term for tech debt, feature debt, design debt and so on. These are continuous improvements to your product that make it more robust, secured and high performing.

For example:

• Improve performance of product

• Eliminate frequent bugs and friction points

• Test automation

So where should you allocate your scarce resources?

This is a bit of an art where your product prioritization really comes into play.

You need to take into account

• Your current revenues (NRR) and customer situation (CSAT)

• Your competitive positioning

• Your product’s stage

• Your resources

• Investment appetite

Let’s start with your product stage. Typically, there are 4 stages to a product or a company.

Early → Growth → Scale → Mature

At each stage of your company, your relative priorities should be different.

Early → Focus more on Bread And Better so that you can sell to more customers that fit your core ICP. At this stage, you don’t want to make any big bets or try something different. Focus is key. While you still have to make sure your product works and performs well, you may be able to wing by a few bugs here and there, unless they are critical.

Growth → Now you are growing to more and more customers of the same type. Product quality will become important as you reach the late adopters who are less forgiving. You will be focused on retention a lot more at this stage so fixing those frequent bugs will be key, while still continuing to add product capabilities for a wider range of customers. This is the most challenging phase for any product team as more and more balls need to be juggled.

Scale → You will continue to do the things you did in growth stages but may want to add some Big Bets. e.g. Grow into a new industry or geo, introduce a new pricing model, create a new add on product.

Mature → If you are a mature product, you would continue to improve the existing product to keep customers satisfied and keep retention high. But you will need to diversify. Take on some big bets to expand. Or do something entirely new.

Here is a helpful summary. This is open to interpretation as each context is different.

Beyond your product stage, there are other factors that will influence your priorities.

Your competitive position in the market. If your position is weak, then you will need to make sure your product works flawlessly to avoid customer churn and stay ahead of the competition. But if you are a leader, then perhaps you can experiment and make some big bets.

You will also need to look into your existing customer situation. If your customer satisfaction is low, then you may want to focus on bread and better and brilliant basics. If customer satisfaction is very good then you can take some big bets or even go after breaking bad items.

Finally, the investment appetite. Say you have convincing proof that doing X will move revenue 30% but requires additional resources. Your company many not have the cash or investment appetite yet. This happens a lot in PE backed companies.

There are a million ways to think and arrange your ideas.

Your discovery, validation, feedback and evidence is going to be used to frame your choices.

And as you choose, you will realize there are trade offs.

Say you focus on Bread and Butter and Brilliant Basics in a given quarter, then management should not expect any new products to sell.

If you focus on Big Bets, then product quality might suffer and support may need to field more tickets.

There is no formula or framework or algorithm at this step. This is the most difficult part of product planning. Your job as a PM is to use your evidence to come up with the best combination of the plan that will help move the revenue needle with the least risk.

Now, your situation may be different. And your product mix might not align with this framework. And thats OK. Just make your own criterion and stick to it. There are no hard and fast rules here.

You will of course need buy in from your management and stakeholders. When you present your plan, make sure to include :

• Customer outcomes
  What are the outcomes that the customer are going to expect?

• Expected business impact
  What are the expected revenue add to the business?

• Tradeoffs
  What are we not doing what is the impact ?

What kind of tradeoffs should be expected?

If you put lower emphasis on :

• Big Bets, then

  • No new products to show to customers.

  • Sales and marketing have to consistently work on messaging for the existing product

• Bread and Better, then

  • Existing customers may complain and may result in churn.

  • No new features may lead to product stagnation

• Breaking Bad, then

  • Then you will need to keep growing within your existing market

• Brilliant Basics, then

  • Quality might suffer.

  • Support may have to field more than usual number of tickets.

Once you have crafted your product plan, use the template below to document your plan.

Use this to get stakeholder as well as management buy in. That way you can be sure that you are not missing any major activities that you should be working on. It is also useful to ensure that all stakeholders are aligned and they adjust their own priorities.

Happy planning.

Subscribe now

Dear PMs and founders,

Wish you all a very Happy Thanksgiving.

There is plenty to thank for this year and wanted to express a sincere thanks to you for reading my substack.

You probably noticed my absence here in the past couple of weeks but I will resume after this holiday weekend.

We have been heads down building ZiplyAI. Last couple of months have been busy on the product side as well as from the GTM side as we try to acquire customers.

Thanks are always slow in the beginning but we can now see the light at the end of the tunnel.

All the posts I have written have been based on my experiences, but now that I am a founder I am reinforcing my own beliefs. Easy to commentate from the outside, but when you are in the weeds things are different for you.

There are many things I am relearning in this journey and I will share them here.

1. Story telling is an art and a critical one.
We have to bring our buyers into the emotional journey, not just throw features at them. They need to feel what you felt when you discovered this problem, why it matters, what changes when it’s solved.

2. Value proposition is very subjective.
Sure, we might claim “4x pipeline.” But here’s the thing—every customer sees value differently. Think about the last car you bought. The dealer sold you on safety ratings and fuel efficiency. But you probably bought it because you loved how it felt to drive, or the color reminded you of your first car. Getting to that value point, the one that actually matters to your prospect—that’s where deals happen.

3. We are in the weeds. We know everything.
But your prospect is seeing this for the first time. You have to guide them. Just yesterday, I met with someone after our first demo. He looked at me and said, “Frankly, I didn’t understand your product and how it will help me.” Ouch. That hurt. But it was a gift—because I realized I was talking at him, not withhim.

4. Prioritization. Oh well.
Much has been written about it. But in early stages, the criteria is different. It’s pure revenue based. Now that’s oversimplified and I’ll explain why, but that’s your filter.

So these are the topics I will be writing about in detail over the next few editions.

If you are new, please subscribe today. Link below.
You can read my articles at www.businessofproducts.com

And share your experiences in the comments below.

I will also resume the Friday GEMs.

Every Tuesday I share my experiences in B2B Products and startups.
Every Friday I share a GEM I found useful.

Subscribe now

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Hi Founders and PMs,

If you’re building a new B2B product from zero, you’ve heard the advice: “Focus on your ICP. Say no to the wrong customers.”

Great advice.

Except when you’re starting out, you have no idea who the “right” customers actually are.

The Theory vs. The Reality

The theory is clean. Define your Ideal Customer Profile or ICP. Target them relentlessly. Disqualify anyone who doesn’t fit.

This makes sense after you’ve found product-market fit. You know what works. You know who pays. You know who churns.

But before that? Your ICP is a hypothesis. And you can’t validate a hypothesis without data or experiences.

The Problem with Being Selective Too Early

When you’re at literal zero - no customers, no revenue, no validation - being selective sounds smart. But it creates a different problem.

How do you know if someone fits your ICP without trying them? The nuances of fit aren’t obvious until you work with real customers.

You think you know. You write down criteria. Company size. Industry. Budget. Team structure.

Then reality shows up and proves you wrong.

What I Learned Building Ziply

Building Ziply from scratch taught me that “wrong” customers aren’t always wrong. Sometimes they’re your best teachers.

Here’s what happens when you let early customers in, even if they don’t perfectly fit your imagined ICP:

1. They validate what actually works

I offered Ziply free to early contacts. Friends, former colleagues, acquaintances who showed interest.

Were they all perfect fits? No.

Did they matter? Absolutely.

They found the bugs. They got confused by flows I thought were obvious. They told me what features were missing and which ones were pointless.

Some of those “test” customers are paying customers today. Without them, we’d still be fixing basic usability problems instead of growing revenue.

2. They reveal the real patterns

We thought we knew our ICP. Mid-size companies. Marketing teams. Budget for content tools.

Turns out, that wasn’t specific enough.

We learned the hard way that companies without a full-time content marketer weren’t ready for us. They wanted the results but didn’t have the process to use the product consistently.

That’s a nuance you can’t spot in theory. You only see it by working with customers who struggle.

3. They extend your runway

Let’s be honest about the money part.

Yes, some early revenue is “bad revenue.” Customers who will churn because they’re not a real fit.

But if you’re bootstrapped or low on funding, even small revenue matters. That extra month of runway might be the difference between figuring out your product and shutting down.

I’m not saying chase bad revenue forever. I’m saying it buys you time to learn what good revenue looks like.

So When Do You Start Saying No?

This is the real question.

You can’t be picky at zero. But you can’t say yes to everyone forever either.

The turning point comes when you start seeing patterns:

• Which customers get value quickly vs. which ones struggle

• Who renews vs. who churns

• What company characteristics predict success

For us, it took about 15-20 customers before the patterns became clear. Before that, every customer felt like a unique case.

Once you see the patterns, you can start being selective. Not because a framework told you to, but because you have evidence.

For Product Managers in Larger Companies

If you’re a product manager launching something new inside an existing company, this applies to you too.

You might have more resources than a founder. But you still need to figure out who actually wants what you’re building.

The same principle holds: your initial customer criteria are guesses. You need real users to validate or change those guesses.

The advantage you have is access to existing customers. Use them. Don’t wait for “perfect fit” customers to volunteer. Go find people willing to try something new and give honest feedback.

The Bottom Line

The advice to focus on your ICP isn’t wrong. It’s just premature.

At the very beginning, your job isn’t to disqualify customers. It’s to learn fast enough to know who to disqualify later.

The real question isn’t whether to disqualify early customers. It’s how long till you can be selective.

Curious how other B2B founders handled this.

Did you entertain some “wrong” customer early on?

I’d love to hear your experience in the comments.

Subscribe now

And that’s basically Robs approach. Get a commitment, or a sale.

Now, I am way oversimplifying this. So let’s hear it directly from Rob.

This video is from his podcast where he articulates his approach.

There is a link to his podcast on the Video description.

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

Who will make money? What types of companies will be successful? Who will be left behind?

And more importantly, will I make money?

This is a fascinating article by Jerry Neumann, an investor.
It’s a bit heavy on economics but the author ties it to historical parallels.

Here is the article.

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

I remember people being very skeptical of computers.

Will it make mistakes? Will it print letters correctly or be as good as a typist?

When people used spreadsheets then, someone would cross check the numbers manually using a calculator.

There were a lot of brouhaha about people losing jobs.
We don’t need that many typists?

But my father used to say - the jobs will become even more specialized and importantly your capacity for more work will increase.

In his engineering office, all drawing were made by hand. Then he got AutoCAD. He trained everyone on it. And now he could do more projects without adding headcount.

And guess what, each of these draftsman who got trained went onto to build businesses in engineering drawing services and employed more people.

A disruption like this lifts peoples wages and provides opportunity for more people to work. In India, many women started getting employment because of computers and the opportunities it created.

And now we have AI. The parallels sound eerily similar.

Jobs will go away.
Is AI really as good as it claims?

Here is my lesson learnt.

All of the above is true ….now.

But these questions will get settled. Soon, you will not be able to talk to your friend without resorting to AI. You think I am kidding….

But what I am most hopeful for? There are so many human challenges that are yet to be solved - hunger, wars, diseases, inequality.

I think AI has the ability to uplift society once again. And I am excited for that.

What about you? What is your take?

AI is good, bad, neither? both?

Share

Recently I was learning about Agentic AI and came across this wonderful GEM.
It was shared by

A senior engineering leader at Google, Antonio Gulli, wrote a book about Agentic AI design patterns. And he shared the entire book in a Google doc.

Here it is :
Agentic Design Patterns: A Hands-On Guide to Building Intelligent Systems

This is useful for all PMs and founders who are planning to build AGentic AI into existing products or trying to build new one in their domain.

Share

I share B2B PM articles every Tuesday and a GEM like this every Friday.
Here are the past Gems.

https://productsos.substack.com/t/gems

In this issue, let’s talk about Jobs to be Done and process mapping.

I was in Scottsdale, Arizona for a weekend getaway. We stayed at this large resort spanning multiple acres. You had to drive from your room to get to the reception area.

On Sunday we were to check out and I was dreading the experience. The traditional check out process at any hotel is archaic and cumbersome.

You walk to the reception. In my case, I would have to drive.
You stand in line, which can be a bit of a wait especially at peak times in popular locations.
When your turn comes, you wait for the staff to go through your charges.
They they will print out the bill. Usually, the printer is in a different locations. At times, they will run out of paper or even ink.
They you pay, sign the receipt.

An investment of 5-20 minutes.

Some hotels have improved the process especially for their frequent guests, but they still require you to download an app and sign in. Or go to the check out area and use a kiosk for self check out.

But this time I had a slightly different experience.

In the morning they sent me an SMS reminding me to check out. And once I leave the room, I should simply reply “out”.

Which I did when we left.

A response came back that I had checked out and a copy of the final bill will be emailed.

Done.

This was truly transformational. No standing in line. No waiting.

In order for this process to work, they needed my email, my phone number and a credit card which they gathered when I checked in. In fact, they had the email and phone when I had made the reservation through a travel site.

They could have eliminated the last step by defaulting the check out to say 11am. In other words, the guest could simply walk out.

Jobs to be done

In products, we use this framework called JTBD or Jobs to be done. In this framework, we map out what the user is trying to accomplish. In my case above, my job was to check out. There are no adjectives attached to the JTBD. e.g. my job is not a seamless check out or a smooth check out.

Just “check out”. How you provide the experience is next.

Other examples of JTBD in B2B could be :

• Get help or support

• Learn about products

• Renew a subscription

• Submit quarter end financials

• Create a new account entry (CRM)

• File taxes on time

In each of these case, the user is trying to do something. We call these functional jobs.

Functional jobs are when your customers try to perform or complete a specific task or solve a specific problem as in the examples above.

But in addition to functional jobs, a user also has a social job and an emotional job.

Social Jobs: When your customers want to look good or gain power or status. These jobs describe how customers want to be perceived by others, for example, be considered a reliable employee who knows how to do their job.

Emotional Jobs: When your customers seek a specific emotional state, such as feeling good or secure. For example, achieving the feeling of job security at one's workplace.

Lets’s imagine you are a PM for a CRM company and your user is a sales person.

What might be their jobs?

• Functional: Close a sale, Do account reviews. Renew customers etc

• Social Job: Become an accomplished sales rep who beats quota all the time and join the elite sales club.

• Emotional Job: By consistently beating sales quota, a sales rep ensures job security and a fulfilling sense of purpose.

As a PM, when you design your product for this persona, you take into account all the jobs to be done and create product functionality accordingly.

The next step in this framework is to highlight everything a user has to do as part of their daily work. The process is laid out from a users perspective.

At each step identify any friction points and areas of improvement.

A helpful tool to use in this case is the “Before and After.”Map the process from a users perspective on how they do the job today. Then identify identify friction points in each of the steps.

Once you figure out the friction points, you can find solutions to eliminate.

Here is a simple illustration of the before and after of the check out process.

You may ask what if there are exceptions. For example, there is a wrong charge in the bill or the guests does not want to use the same credit card. Of course, the guests can go back to reception or email or call them.

But how often does that happen? If exceptions only happen rarely, then the processes should not be optimized for those scenarios. Rather, optimize them for the 80% of cases. Exceptions are always there in business and you can create processes around those.

Let me share a real life example of a product I launched a few years ago.

A sales person gets a cold lead via email. They have no other information about this lead. So they do some research about the person. Using Google, or other tools they try to find information about this person and the company. That way, they can have a meaningful conversation. The friction in this process is :

• Takes too long to find information

• Cannot access certain information

These sales people are on the clock as they have an SLA to respond to this lead soon, sometimes within minutes. The ideal solution was to automate the fact finding exercise so that the sales rep can have a meaningful conversation with the lead. This helps to improve the sales call and possibly reduce the sales cycle.

We created a product that gathered information from the external sources automatically. We partnered with other data providers who could also provide more information about individuals and companies. The product then packaged the information nicely and aided the conversation of the sales rep. So the cold lead is not a warm lead anymore.

There is even more opportunity to improve the process. What if AI could create talking points based on the information gathered. e.g. “based on recent news, the company where this prospect works expanded operations Europe and may be impacted by data privacy. Consider talking about the new data privacy check up tool”.

Jobs-to-be-Done (JTBD) has always been about clarity: customers don’t buy products, they “hire” them to make progress in a specific situation. It’s a way of stripping away features, personas, and buzzwords to focus on the outcome a buyer truly values.

Now, as we start designing with agentic AI - systems that can reason, act, and adapt—the JTBD lens becomes even more relevant. Instead of asking, what can this AI do?, the more powerful question is, what job should this AI be hired to do?

Conclusion

JTBD reminds us that customers hire solutions for progress, not features. Agentic AI raises the stakes: if we design agents around tasks alone, they’ll end up brittle and underwhelming. But if we root their design in the jobs customers actually need done—researching, qualifying, drafting, deciding—we create agents that act as true partners.

JTBD provides the why; AI delivers the how. Together, they let us build systems that don’t just automate, but anticipate. The future of agentic AI isn’t about smarter agents—it’s about agents that know which job they’ve been hired to do.

As a product manager, our goal is to transform our users life and allow them to perform their tasks with the least effort. It is imperative to understand the jobs to be done of your user personas and map the before/after processes.

Remember, that users want to solve their pain points but only if they are effortless and friction free.

Your solution may evolve. But JTBD rarely changes.

Subscribe now