Why I Tweaked the Lean Canvas to Create the Paid Pilot CanvasWhat two years of working with startups taught me about moving from business-model hypotheses to revenue-first validation
I Started Questioning the SequenceOver the past two-plus years, through my work with startups and a few incubators, I found myself increasingly questioning a familiar startup sequence: Build → launch → find customers → learn what they want. The teams I worked with were often very capable of building. They could prototype quickly. They could use no-code tools, AI, and existing platforms to get something into users' hands. They could explain the vision. But beneath all that capability, one uncertainty kept coming back:
That question gradually pushed me toward what I began calling revenue-first validation. Not as a new theory I arrived at overnight. It emerged through practice - by working with startups, trying different approaches, seeing what generated useful signals, and noticing what happened when we put an offer in front of a real customer before committing to a larger build. The sequence started to look different: Find a customer → sell an outcome → deliver it → learn → then decide what to build. The more I experimented with this approach, the more convinced I became that sometimes revenue shouldn’t be the thing we wait for after validation. It can be part of how we validate in the first place. Then I Started Seeing the Pattern EverywhereOnce I started exploring this idea more deliberately, I wanted to know: Is this just a tactic that works in a few situations, or is there a broader pattern? So I looked at startups across different domains and business models that had generated revenue before their eventual product was fully built. The examples were surprisingly diverse. Some started by delivering manually what would later become software. Some pre-sold. Some used deposits or early commitments. Some operated behind the scenes with people and simple tools before automating the process. Some tested a new offer with a small group of customers before investing heavily in the eventual product. The specific mechanics varied, but the pattern was remarkably consistent. They didn’t always start by building the thing they ultimately wanted to build. They found a way to create and sell value first. My research surfaced examples ranging from Canva’s early manual design services and ClassPass’s manual booking model to Peloton’s pre-orders, Urban Company’s manually coordinated services, Dunzo’s early task fulfilment, Mailchimp’s early paid model, and Uber’s manually coordinated rides.
That changed the way I thought about validation. The question wasn’t always: “Can we validate the product before building it?” Sometimes the more useful question was: “Can we sell and deliver the value before building the product that eventually scales it?” That became an important distinction for me. So I Started Experimenting With Revenue FirstResearch gave me the pattern. Working with startups gave me a place to test it. Over the past couple of years, I experimented with different ways of getting startups closer to revenue before making a larger product commitment. Sometimes it meant pre-selling. Sometimes it meant a concierge approach. Sometimes the “product” was actually a manually delivered service. Sometimes it was a combination of software and human effort. Sometimes the experiment was built around a very specific outcome for a very narrow customer segment. Sometimes the biggest variable we were testing was price. There wasn’t one formula. And that was actually the point. I wasn’t looking for another validation trick. I was trying to understand something more fundamental:
The more I experimented, the more I realized that the answer wasn’t necessarily an MVP. Sometimes it was much smaller. Sometimes it was much more manual. And sometimes it looked nothing like the eventual product. But it had one important characteristic: There was a real customer on the other side of it. I Realized “Validation” Was Too Broad a WordWe often tell founders: “Go validate the idea.” But after doing this work repeatedly, I started to find that phrase almost too vague to be useful. What exactly are you trying to validate? You might be testing whether:
These are very different questions. A customer saying:
is a signal. A customer saying:
is another signal. But neither necessarily tells you what happens when you ask:
And even payment alone isn’t enough. Once someone pays, another set of questions appears:
This is what increasingly attracted me to revenue-first validation. It doesn’t mean that every assumption has to be tested through payment. It means that when the question is commercial, a real transaction can create a much richer learning environment than interest or intention alone. The customer isn’t simply giving you an opinion. They’re committing. And now you have something much more concrete to learn from. Then I Had to Reconcile This With Lean CanvasWe’ve all used the Lean Canvas as a powerful way to make the business model visible and to expose the underlying hypotheses.
In the Lean Canvas, you step back and think about the business as a whole:
And importantly, you don’t treat those answers as facts. They’re hypotheses. You question them. You identify the assumptions that matter most. You ask: Which assumption is weakest? What experiment could test it? That logic makes complete sense to me. But when it comes to designing experiments, the Lean Canvas did not aid in articulating experiment design and, more particularly, it did not necessarily nudge towards a commercial validation, leaving open doors to other things such as customer interviews for pain point validation. And anything other than a “commercially” focused validation has been proving less and less attractive in building a base for a new startup. I Needed an Experiment-Design Canvas.The Lean Canvas helps me reason about the business hypothesis. The Paid Pilot Canvas helps me design the paid experiment I want to run against that hypothesis. That’s a subtle difference, but it became increasingly important in practice. Once you’ve decided that a paid pilot is the experiment you want to run, a whole new set of questions appears. You have to get much more specific:
These aren’t primarily questions about the eventual business model. They’re questions about the immediate commercial experiment. That distinction is what led me to the Paid Pilot Canvas. Lean Canvas helps you see the business you are hypothesizing. Paid Pilot Canvas helps you design the experiment you are about to put into the market.
The ShiftThe more I worked this way, the clearer the distinction became. The Lean Canvas makes you zoom out. You look at the potential business as a whole. You think through the different parts of the model and the assumptions connecting them. The Paid Pilot Canvas makes you zoom in. For a moment, you forget about the entire business. You focus on one experiment.
This is what I mean by micro-designing the experiment.
And that distinction matters. Because an experiment can fail simply because it was poorly designed. The customer may be too broad. The outcome may be vague. The price may not match the value. The pilot may try to solve too much. The delivery model may be unnecessarily complicated. Or you may finish the pilot without knowing what you actually learned. The canvas forces those decisions to be made before you enter the market. Why “Paid” MattersThis is also where the word paid became important to me. Because payment changes the experiment. It brings commercial commitment into the test. Instead of only asking whether someone likes the problem or finds the solution interesting, you can start learning about: Problem + urgency + value + willingness to pay + price + delivery + outcome. The customer has to make a choice. Not simply express an opinion. That distinction became particularly interesting when I looked back at the examples I had researched. Across different businesses, early revenue took different forms - re-orders, deposits, paid services, manually delivered offerings, and other forms of early customer commitment. The mechanics differed, but the underlying idea was similar: get closer to a real commercial exchange before making the larger investment. And I think this matters even more now. Building is getting cheaper. AI, no-code tools, APIs, and increasingly accessible development infrastructure make it possible to create software and prototypes faster than ever. That’s fantastic. But it creates a new risk. We can build something impressive before we’ve established that it deserves to exist. The cost of building has fallen. The cost of being wrong hasn’t. That is why I’m increasingly interested in revenue-first validation. The Paid Pilot Canvas Emerged From the PracticeI didn’t start this work thinking: “I’m going to invent a new canvas.” The canvas came much later. First came the work. Then the experiments. Then the recurring questions. Then the recurring mistakes. I kept finding myself helping founders think through the same things: Who exactly is the pilot for? What are we promising? What is the customer actually paying for? How much should we charge? What can we deliver manually? What should we measure? What should we learn? And perhaps most importantly: What should we not build yet? After doing this enough times, I realized these questions could be brought together on a single page. That was the origin of the Paid Pilot Canvas. The intention was similar to what I appreciated about the Lean Canvas:
But the object being designed was different. The Lean Canvas helps you visualize your business-model hypothesis. The Paid Pilot Canvas helps you visualize the design of the immediate paid experiment. And I think that distinction is important. The canvas isn’t the methodology. It’s the compression of the methodology into a practical tool. The real methodology is what happens around it: Design → Sell → Deliver → Measure → Learn The canvas simply gives you a place to think before you begin. The Canvas Isn’t the End. It’s the Beginning.This is perhaps the most important thing I learned through using the approach. Completing the canvas doesn’t validate anything. The market does. The canvas helps you design the experiment. Then you have to take it into the world. You find the customer. You sell the pilot. You deliver the promised outcome - even if parts of the delivery are manual. You measure what happened. And then you learn. That creates a loop: Design → Sell → Deliver → Measure → Learn → AdaptAnd the learning can take you in very different directions.
All of those are valuable outcomes. Because the purpose of a pilot isn’t to prove that your original hypothesis was right. It’s to make the next decision better. After each pilot, you can ask:
That’s when the Paid Pilot becomes more than a document. It becomes a learning loop between the business hypothesis and the market. The Deeper Belief: Earn the Right to Build MoreAfter working this way for the past couple of years, I’ve become increasingly convinced of something:
Not because building is inherently wrong. And not because every startup needs to operate manually forever. But because every additional investment in product, technology, and infrastructure is a bet. And I’d rather make that bet with evidence than assumption. Don’t build more simply because the hypothesis makes sense. Build because your experiments have given you a reason to. That’s what I now see as the real role of a paid pilot. It isn’t designed to prove that the entire business works. It isn’t supposed to eliminate uncertainty. It is supposed to generate enough evidence to justify the next investment. Sometimes that next investment is another pilot. Sometimes it’s a better offer. Sometimes it’s a new customer segment. Sometimes it’s automation. And sometimes it’s finally time to build the product. Why I Created the Paid Pilot CanvasLooking back, the Paid Pilot Canvas wasn’t really the starting point. It was the result of the journey. I started with the tools I already had. I worked with startups. I became increasingly interested in revenue-first validation. I researched examples across different industries and business models. I experimented with different ways of running paid experiments. And eventually, I found myself needing a simple way to design the experiment before executing it. That’s what the Paid Pilot Canvas became. Not a replacement for Lean Canvas. Not another business plan. Not another framework to fill out and forget. A one-page way to think through the smallest commercial experiment you can take to a real customer. The distinction, for me, is now quite simple: Lean Canvas helps you visualize the business you might build. Paid Pilot Canvas helps you design the experiment that tells you what deserves to be built next. And sometimes, the most valuable thing you can build first isn’t the product. It’s the evidence. The Paid Pilot CanvasI’ve captured this approach in a one-page Paid Pilot Canvas for founders who are exploring an idea, building an MVP, or testing a new offer. If you’re trying to figure out what to build next, perhaps the better first question is: What can I sell and deliver first to learn whether it’s worth building?
|
