None of the product management trainings or experts teach you about growth.

At least not the right way to think about it.

Cast your mind back ten years. Blue Apron was one of the hottest Silicon Valley startups. It popularized the idea of meal kits, alongside companies like HelloFresh in Europe.

The promise was simple. Skip the trip to the grocery store and the nightly “What’s for dinner?” debate. Blue Apron delivered a recipe and every pre-portioned ingredient you needed, right to your doorstep, so you could cook a fresh meal at home.

The business exploded in growth. By 2016, it reported $795 million in revenue. It IPO’d in 2017 at nearly $2 billion. It had over 40% market share. You can see why everyone got so excited.

The message in the S-1 was simple: We’re a rocket ship and to win we just need to do more of what we are doing: grow customers, grow orders.

Take a close look at how they talked about the business:

They clearly celebrate Orders and Customers. Repeat Orders were disclosed too, but almost as an afterthought — a “useful metric”. Not the one investors were encouraged to focus on.

So how did Blue Apron grow those headline numbers? By making the first purchase almost impossible to refuse.

New customers could get their first box for about half price. A two-person box with three recipes and six servings normally cost around $60, so your first order was roughly $30 off.

It felt like an incredible deal. You’d be hard-pressed to buy the same ingredients at a grocery store for less.

The problem came after that initial deal.

Once customers rolled onto the regular subscription price, Blue Apron stopped feeling like a bargain and started feeling like an expensive way to have groceries delivered. On top of that, it inherited all the headaches of a physical fulfillment business: late deliveries, damaged ingredients, warm food, missing items, and ruined dinner plans.

Not surprisingly, many customers didn’t stick around. On average, they placed only about 4 orders before churning.

The economics started to look ugly.

In 2016, Blue Apron’s gross margin was about 33%. That’s roughly $20 of gross profit on a $60 box before marketing and overhead. (That may not be the exact contribution margin, but it’s close enough to understand the business.)

The problem was they were spending $94 in marketing to acquire a each new customer.

Think about that for a second.

$94 to acquire a customer who was often motivated by cheap groceries and the novelty of trying something new.

In 2017, average revenue per customer was $245. At a 33% gross margin, that’s only about $75 of gross profit before marketing and overhead. Against acquisition cost.

And that’s before paying for fulfillment centers, customer support, technology, G&A expenses, credits and refunds, and depreciation.

Even a 7th grader can tell you the math was absolutely horrible.

If you’re spending roughly $94 to acquire a customer who generates only about $75 of gross profit before operating expenses, there’s very little left for shareholders.

But as long as the marketing engine kept running, the company could continue reporting growth in customers, orders, and revenue, along with the familiar promise that profits would come once the business reached sufficient scale.

But they were growing the wrong thing.

Total orders and total customers were not the atomic unit of growth for this business.

The thing they needed to grow was repeat subscription customers.

Those were the customers who built a lasting habit. They enjoyed discovering new recipes, avoided the hassle of grocery shopping, and valued having exactly the right ingredients delivered each week.

The problem was that this group was much smaller than the group willing to sign up for a couple of weeks simply because the heavily discounted first box was cheaper than a bag of groceries and the novelty cooking experience.

Total orders and total customers were a smokescreen.

In reality, the company had built an expensive fulfillment network designed to serve a massive volume of customers who had little intention of becoming long-term subscribers.

That’s a fundamentally flawed business. And it showed up in the financials:

And once Blue Apron became a public company, investors looked past the headline growth metrics, started asking harder questions, and crushed the stock.

When marketing spend inevitably came down under public market scrutiny, the underlying economics were exposed. Customer acquisition slowed, order volume fell, and the growth story unraveled.

After a disastrous run as a public company, Blue Apron was eventually acquired in a deal valuing the company at about $103 million — roughly 95% below its IPO valuation. By that point, it had burned through around $700 million in equity capital.

Too bad, because there was probably a solid business hiding underneath it all.

The winning strategy wasn’t to acquire as many customers as possible. It was to find the people who genuinely wanted meal kits to become part of their weekly routine and build the business around them.

Those customers may have been more expensive to acquire, but they also would have stayed longer, spent more over their lifetime, and justified a leaner, more efficient operation.

In other words, they needed to optimize for a completely different metric. Not total customers or total orders. But repeat customers who formed a lasting subscription habit.

That may not have produced the same eye-popping growth charts, or even the same IPO valuation. But it would have given the company something far more valuable: a realistic path to becoming a durable, profitable business.

Like I said, PM trainings and experts don’t teach us to think about growth the right way.

They focus on PLG, “pirate metrics,” activation and onboarding, or some vague notion of a “north star” metric, assuming these will automatically lead to sustainable growth outcomes.

Finance isn’t much better.

Open almost any finance textbook and you’ll find a growth tree. Profit breaks down into revenue and cost. Revenue breaks down into price and volume. Volume breaks down into customers and purchases. Keep drilling down and eventually you end up with a long list of metrics that all appear equally important.

It’s not that they’re all wrong. It’s that they quietly lead you to the wrong place.

Two problems. First, it encourages you to work from the top down, giving every number equal billing, treating each branch as if it deserves equal attention.

Second, they ignore the interdependency between the numbers. They influence each other. And somewhere in that web of relationships, there's usually one variable doing far more work than all the others.

Run Blue Apron through it, and you only get to the problem when you look at the relationships between the branches.

On paper, almost everything appeared to be moving in the right direction.

Revenue was growing.
Volume was growing.
New orders were growing.

But…

New orders weren’t becoming repeat ones often enough.

They had built a wonderful looking yacht with a hole at the bottom.

While the headline numbers looked fantastic, the business was quietly taking on water.

If management had instead obsessed over converting first-time buyers into long-term subscribers — reducing early churn and increasing renewal rates during those critical first few months — they almost certainly would have built a smaller business at first. But also one that would have been far more durable.

Every business has a core unit of growth that drives its business model.

It’s the smallest unit of customer behavior that, when it increases, reliably makes the business more valuable over the long term. Grow the right unit, and revenue and profit tend to follow. Grow the wrong one, and you can end up with impressive-looking growth built on weak economics.

Blue Apron treated total orders as its unit driver of growth. Instead, it should have been something like customers with three months of uninterrupted orders. That’s the point where someone stops being a bargain hunter trying a discounted meal kit and starts becoming the kind of customer the business was actually built to serve.

Some other examples:

Of course, the hard part is figuring out what your atomic unit of growth actually is. Once you’ve identified it, everything else starts to fall into place. It becomes your true North Star. Product strategy, roadmap decisions, experiments, and investments can all be evaluated against a single question:

Will this increase our atomic unit of growth?

That’s powerful because growing the right unit doesn’t just improve one metric. It creates a cascade of second- and third-order effects that strengthen the business over time.

Imagine if Airbnb had decided its goal was to maximize listings instead of nights booked. Marketing dollars would have flowed toward recruiting more hosts. Engineering would have prioritized host acquisition. Supply would have exploded.

But if demand didn’t keep pace, occupancy would fall, host earnings would decline, the guest experience would suffer, and the marketplace would slowly weaken — even while the headline metric of total listings kept climbing.

That’s why one of the most valuable things you can do as a product leader is help your business identify the right atomic unit of growth.

So where do you start?

Why, with the customer’s problem, of course! Exactly the thing we product people focus on day in and day out!

The best atomic units of growth are based on the job your customer is hiring you to do — not the thing you sell.

The best atomic units of growth measure how often you solve an important customer problem in a way that customers want to come back and repeat.

When you get that right, something powerful happens: your success and your customers’ success become tightly aligned.

And every time you help customers solve their problem again, you create value for them and, in turn, make your business more valuable too.

Once you’ve got a candidate atomic unit of growth, run it through these four tests:

This isn’t a perfect science. You may have to experiment. The unit that moves earliest is often the hardest to prove causally. The one closest to real activity might be the messiest to attribute.

When you find it, everything falls into place. If it doesn’t feel quite right, it probably isn’t the right one.

Large companies often have multiple businesses, products, or business models under one roof. Each may have its own atomic unit of growth, and that’s perfectly ok.

The warning sign is when what should be one business is behaving like several disconnected ones.

Great businesses have flywheels. Growing one unit makes the next unit easier, cheaper, or more valuable to grow. If your atomic units don’t reinforce one another, you don’t have a flywheel. You have competing growth engines.

And competing growth engines almost always make growth slower, more expensive, and less durable than it needs to be.

So, with an atomic unit of growth defined, the crucial dynamic becomes, of course, growing the numbers of atoms quickly, efficiently and profitably.

I think about this in ten flavors:

Capacity Growth — building more features — is used to fuel growth far more often than it should. Product teams love to pump out features. Heck, even tech investors get excited about cool new features, and use them to artificially bump valuations.

But while building more features is fun and cool, it doesn’t guarantee growth. The key question is whether the new capability makes the atomic unit of growth cheaper, stickier, or more valuable.

There are seven types of features:

PMs love to focus on #7. Engineering teams love to focus on #5, #6, and #7. The theory is that:

Feature ➜ Outcome

Instead, savvy product leaders think like this:

Feature ➜ Atomic Unit of Growth ➜ Outcome

In other words, the feature itself doesn’t create value — it changes the behavior of the atomic unit.

You can think about each of feature types in relation to the atomic unit of growth:

So when you look at any feature as a source of growth, get to the crux of it. Which flavor of growth does it actually buy you?

Because — and I’m sure I don’t need to explain to you — that a feature always comes with baggage, called technical debt. And, as a product person, you well know that tech debt can quietly drag your growth.

That’s the depth you need to get to before you weigh a feature against some other organic alternative.

Now, let’s make this tactical for you:

Take one product, segment, or use case you’re trying to grow.

1. What are you currently trying to grow?

Write down the metric everyone talks about — revenue, customers, accounts, orders, users, locations, downloads, pipeline — whatever it is.

2. What is the customer actually paying you to solve?

Not what you ship. Not your product. The problem or or desired outcome.

3. What is the smallest repeatable unit of that value?

This is your candidate atomic unit of growth. For example: a completed job, a renewed member, a paid seat, a night booked, a repeat dinner occasion, a flying hour, a mile driven, etc.

4. Does it pass the four tests above?

Is it causal, not correlated? Atomic, not aggregated? Leading, not lagging? A lever, not a thermometer?

Then finish this sentence:

Too often products grow the wrong thing. The unlock is finding your true atomic unit of growth — the thing your customer actually came to you to solve. Grow that, and profit follows, because you’ve tied your success directly to theirs.

The atomic unit of growth is just one example of a bigger shift.

Most product leaders are taught how to build products.

Very few are taught how to build businesses.

That’s why so many roadmaps optimize for outputs instead of economics, and why so many product teams struggle to explain how their work will actually create long-term shareholder value.

The Business Fluency for Product Leaders Masterclass is designed to close that gap.

Over two live days, Mike Smart and I teach experienced product leaders how to think like executive teams do: capital allocation, revenue models, unit economics, financial strategy, and how to turn product initiatives into investable business decisions.

If you want to become the product leader who understands not just what to build, but why it creates value, I’d love to have you join us.

The next cohort begins in July and space is limited. If you’d like to build those skills, we’d love to have you join us.

That’s it for today.

Have a joyful week, and, if you can, make it joyful for someone else too.

cheers,
shardul

Shardul Mehta
I ❤️ product managers.

​Follow me on LinkedIn