Dead on Arrival: The Launches That Force a Founder’s Hand

Here’s a quote that’s been stuck in my head all week: "Most commercialization failures are not created at launch. They are merely revealed there." Oof. Yeah. That lands. That’s from that Knox Institute paper, right? It feels like it should be tattooed on the inside of every founder’s eyelids. Exactly. And it’s the perfect bridge from our conversation last week about brutal pivots. We talked about the aftermath of a failed plan, but today I wanna talk about the moment before the pivot—the launch that forces your hand. The one that was dead on arrival and you just didn't know it yet. Oh, I know the one you mean. You’re talking about Fetch. The company that’s now processing, what, 13 million receipts a day?

Thirteen million a day. A reported seven-hundred-million-dollar gross revenue run rate. Profitable. They are a monster. But their first product was a complete, total, catastrophic failure. And that's what makes it such a good story. Because they didn't just stumble. They built this incredibly complicated, elegant, and fundamentally doomed machine. They called it the "three-legged stool." Which is already a red flag. Anytime your business model needs three different, independent groups to all do something perfectly in sync… you’re in trouble. So break it down. What were the three legs? Okay, so leg one: the shopper. They go to the grocery store, they want rewards. Simple enough. Leg two: Fetch, the company providing the rewards platform.

Fine. Leg three… this was the killer. The grocery retailer itself. Ah. So they needed the store’s permission. Permission, integration, everything. The original idea was a checkout-based system. It had to plug directly into the retailer’s point-of-sale software. So to launch in ANY city, the Fetch sales team had to go store by store, chain by chain, and convince them to install their software and train their cashiers. Oh, that’s a nightmare. That’s not a tech company, that’s a door-to-door sales company with a cripplingly long sales cycle. I mean, can you imagine the pitch? "Hi, we’re a no-name startup. Please inject our code into the single most critical piece of technology in your entire business—the thing that takes the money." Right?

"And then train your entire, very-high-turnover staff on how to use it. For free!" It’s insane in retrospect. But you can see how they got there, right? It feels more… legitimate. More defensible. A deeper moat. It feels like a 'real' business partnership. But you’re not building a product, you’re building a series of one-off consulting gigs. Your growth is capped by the number of salespeople you can hire and how many flights they can take. It’s the opposite of scalable. And that’s what the data shows. CB Insights says 42% of startups fail because of no market demand. But a quote I saw from Userpilot reframed it. They said often the market need wasn’t absent, it was just… misunderstood.

People did want easier savings. Fetch’s CEO, Wes Schroll, even said that. He said, "The first thing to fail… was not the premise that people wanted easier savings. It was the route to them." The route to them. That’s it. They put a mountain range between the customer and the value. And every single grocery store was another mountain they had to convince to let them pass. So what does that feel like, inside the company? You’ve raised money, you’ve hired people. You’re all in on this vision of this beautiful, integrated system. But the numbers aren’t moving. It’s denial, at first. You know? You blame the execution. "Oh, we just need better salespeople." Or, "This one retailer is just slow, the next one will be faster." You look for any reason other than the core assumption being wrong.

And the core assumption here was that retailers would be willing partners in this. That the value for them was clear enough to overcome the massive friction of implementation. Which it never is. What was the value prop for the retailer? Vague promises of customer loyalty? More data? They already have loyalty programs. You’re asking them to do all the work for a benefit they probably feel they can get themselves. So you’re burning cash. The second biggest reason startups fail, right? 82% fail because of cash flow problems. And Fetch is spending all this money on sales teams, on custom integration work… and they’re getting, what, a handful of stores live? It’s a leaky bucket. A firehose, really.

And the whole time, the clock is ticking. You can just feel the runway getting shorter and shorter. And the conversations in the boardroom must have been brutal. The investors are looking at the burn rate versus the adoption rate and the math just isn't mathing. Right. And this is where that Knox Institute quote comes back in. The failure was baked in months, maybe years, before launch day. It was baked in the moment the whiteboard diagram had "Retailer POS Integration" as a required box. The launch just held a mirror up to the bad decision. So then what happens? The moment of truth. Do you double down, or do you find what they called "the escape hatch"? The escape hatch. I love that phrasing.

It wasn’t a "pivot." It was an emergency exit. And it was so brilliantly simple. They asked: what is the one thing that EVERY shopper gets from EVERY store, that requires zero permission and zero integration? A paper receipt. A paper receipt. A little trail of data that the store just gives away. So they scrapped the entire three-legged stool. They threw out the point-of-sale integration, fired the sales team probably, and built an app with one function: take a picture of your receipt. And in that one move, they took back all the power. They no longer had to ask for permission. Their growth was no longer tied to a retailer’s IT department schedule. It was only tied to how many people they could convince to download an app and take a picture.

A waaaay easier problem to solve. A marketing problem, not an impossible sales and engineering problem. And look what happened. 13 million receipts a day. They found a way to get the data they needed without ever setting foot in the store. There’s a quote from the analysis that just nails it. "The winning product did less at checkout and learned more after it." Say that again. That’s the whole lesson. "The winning product did less at checkout and learned more after it." Their first product tried to do everything at the point of sale. It was complex, it was integrated, it was a heavy lift. The winning product did literally nothing at checkout. It was invisible. The work happened later, on the user's couch, in five seconds.

And that’s what this means if you’re a founder listening to this. You have to ask: where is the friction in my model? Who has to say 'yes' for me to grow? If the answer is a long list of people who are not your customer, your model is probably broken. You have to find the path of least resistance. Fetch leveraged a behavior that already existed—people get receipts. They didn't have to teach anyone anything new. They just attached a reward to a thing people were already doing. So let's pinpoint the moment of failure. It wasn't when the first grocery store said no. It wasn't when they missed their first quarter's target. No. The moment of failure was on day ONE. Before they wrote a line of code.

It was the moment someone stood at a whiteboard and drew a box that said "Integrate with Retailer" and everyone in the room nodded. Yes! That's it. That was the original sin. It was the untested, unvalidated assumption that this was a feasible path to scale. Everything that came after—the slow growth, the cash burn, the painful sales cycles—it was all just a symptom of that one, core, flawed belief. And it's so seductive. As a founder, you fall in love with the elegant, complex solution. It feels more defensible, harder to copy. You think a complex problem needs a complex answer. But the market almost always rewards the simple answer. The one that removes a step, not the one that adds three.

The first Fetch product added steps for the retailer, for the cashier, for everyone. The second Fetch product removed the need for all of them. It put the user in complete control. Scan or don’t scan. The business's success or failure rested entirely on the value proposition to that one person, not on a delicate negotiation between three different parties. So the lesson here is almost painfully simple. Don't build a business that relies on other people's permission to grow. Find the "paper receipt" in your industry. Find that trail of data or that existing customer behavior that you can latch onto without needing to ask for a key to the building. And you have to be brutal about your assumptions.

What is the one single belief that, if wrong, will sink your entire company? For Fetch, it was the belief that retailers would be willing partners. And they were wrong. They were catastrophically wrong. But they were smart enough to realize it and find the escape hatch before the plane hit the ground. A lot of founders aren't. They just keep flying, insisting the runway will appear. They ride the bad assumption all the way down. Because they’ve tied their ego to it. They’ve told their investors it’s the right plan. Admitting it’s wrong feels like admitting they were wrong. Which they were! And that’s okay! I mean, that’s the whole point of this show, right? Being wrong is part of the process.

The winners are just the ones who admit it fastest. Okay, so let's bring this home for the founder who's maybe feeling a little queasy right now, listening to this. Maybe their whiteboard drawing has a few too many "three-legged stool" components. Yeah, they're looking at their own beautiful, complex machine and starting to sweat. What’s the first thing you do? If you suspect your core assumption might be flawed, what is the action you take tomorrow morning? You have to design the cheapest, fastest possible test for that one assumption. Forget the full product. Forget the elegant UI. What is the absolute minimum thing you can do to prove or disprove that belief? For Fetch, they could have literally just gone to a mall and offered people five bucks to show them their receipts.

Like, a concierge MVP. Just you, a human, faking the entire backend of the system. Exactly. Before you build a single line of integration code, can you manually prove that retailers are even willing to have a conversation? Can you get ten of them on the phone? If you can't get ten on the phone, you are not going to get a thousand to install your software. It’s about de-risking the single biggest point of failure. And I think founders, especially technical founders, often miss this. They de-risk the tech. "Can we build it?" Yes, you can almost always build it. That's not the question. The question is, "Should we build it?" And even more, "If we build it, who has to give us permission for anyone to even use it?" Mm-hm.

And the Fetch story is the perfect example of the stakes. They had the right idea—people want rewards. They had the right market—grocery shoppers. They had the wrong mechanism. And that one detail was the difference between bankruptcy and a company with a seven-hundred-million-dollar revenue run rate. It's staggering when you put it like that. The entire outcome of the company hinged on changing the method of data collection from "ask permission" to "leverage what's already there." It all goes back to that Knox quote. The failure was there all along. The launch just made it impossible to ignore. The market is the ultimate truth-teller. It doesn't care how cool your tech is or how smart your team is.

It just cares if you solve a problem with the least amount of friction possible. So the final thought for anyone listening is this: look at your business model. Find the friction. Find the person you’re depending on who isn’t your customer. That’s your point of failure. And your job isn’t to convince them. Your job is to design a new model that makes them completely irrelevant.