Repair at Scale: What Upkept Taught Us About Building the Right Kind of Business

There's a moment early in this episode of HappyPorch Radio where Robin Atkinson, co-founder of the clothing repair platform Upkept, says something that sounds almost throwaway, but ends up being the key to everything that follows.

"There is no such thing as unskilled labour," she tells Barry, "and professional sewing is as hard as welding."

It's the kind of line that could sit on a poster. But spend some time with Robin and her husband and co-founder Cash Acrey, and you realise it isn't a slogan at all. It's the operating principle behind a company that started with a question about jobs, ended up solving a problem about clothes, and along the way built its own software because nothing else on the market came close to fitting what they actually needed.

The insight nobody was looking for

Robin didn't set out to build a repair company. She spent nine years running workforce development programmes in the sewn trades, trying to work out how to build meaningful sewing jobs in America. The people she was training, though, kept telling her something she hadn't expected: they didn't want to spend their days on a manufacturing line, sewing the same seam ten thousand times.

"Mass production is more like engineering than it is like art," she explains. "If you're going to sew 20,000 units of something, this is not an art problem, this is an engineering problem." The people coming into sewing in the US, by contrast, tend to have a haptic, problem-solving mindset. They wanted the kind of work where every project is similar, but never quite the same.

That's when Robin went home one evening and, mostly for fun, modelled the financials on what repair might look like as a business. She turned the laptop round to Cash, a finance professor, and said maybe they should look at this properly. The rest, as Barry puts it, is a company built on the idea that the technology isn't the product. Repair is the product. The software just needs to get out of the way.

Barry picked up on exactly this in conversation: "One thing that's interesting…is that you're coming at that from the skill set, from the job, from the workforce point of view, rather than from repair specifically." It's a subtle distinction, but it explains why Upkept doesn't is not a marketplace connecting customers to freelance seamstresses. It's a centralised operation in Northwest Arkansas, built from the ground up, mailing clothes in and out the way you'd order a delivery.

Craft that scales, without becoming a factory

The tension Upkept has to balance every day is that every repair is genuinely one of a kind, while the business itself needs to run at real scale to make the economics work. Robin describes niching down within the warehouse so that some people become denim experts, others become zipper specialists on outdoor gear, all in service of efficiency. But she's careful to draw the line before it tips into treating people like parts on a line.

Cash puts it well: "We don't just want all hyper-specialised, like, insects. Some people have a fundamental strength and skill set, and so there are bricks, and then some people are grout. Some people want to go where the gaps are and just grab a problem and fix it."

Robin calls the whole balancing act "an orchestra problem." Push on one part of the operation and something else will slip, so someone always has to be watching the whole picture. It's a useful reminder that scaling a service business built on genuine craft is never a solved problem. It's a moving target you keep adjusting.

That honesty carries through into how Upkept talks about money, too. Robin is refreshingly direct about turning down the wrong kind of investment, the kind that would ask them to cut headcount and squeeze margins in the name of growth. "If we compromise early, we actually shrink the size of the opportunity too," Cash adds. Staying true to the model, rather than chasing the fastest possible scale, is what keeps the market they're trying to build from collapsing back into a smaller, less useful one. Previous HappyPorch Radio guests have echoed this sentiment as well. 

Barry, who's currently making his way through Eric Ries's book Incorruptible, drew out something that sits close to the heart of what this podcast keeps returning to: when growth and mission are genuinely aligned, growth stops being something you have to defend against your values and becomes proof that they're working.

Building the tool because the tool didn't exist

The second thread running through the conversation is how Upkept's platform came to exist at all. Cash had managed development teams before, written specs, signed off on projects. What he hadn't done was write the code himself, until AI coding tools made that suddenly possible.

He's candid about what that was actually like in practice. It wasn't magic. 

"I found that when I started, I was arguing with the bots quite a bit. I had to wrangle them a lot," he says. His approach was to check the work from every angle he could, the interface, the console, the database, so that nothing could slip through unnoticed. Slowly, with better processes, automated testing and proper documentation, it got smoother. As he puts it, "it's maturing. It's growing up. My baby's growing up."

What's notable is why they built rather than bought. They looked at existing platforms and could see ways to bolt something together, but every option meant managing someone else's system that was never designed for what they needed, and layering on customisation until it became its own kind of burden. Building it themselves, on a shoestring, gave them something that fit precisely and that they could keep shaping as the business changed. Cash is clear that this would have been a far more expensive undertaking even a few years ago. The tools have changed what's possible for a founder with the right instincts and enough persistence to hold the AI to account.

While Cash fortunately had the skills and persistence to shape something that works for Upkept, there’s another story we’ve often seen elsewhere: a founder gets a huge amount of the way there using AI coding tools, and the value they've created is real. But there's a point, whether it's security, scale, integration with a partner's systems, or simply the accumulated weight of decisions made quickly under pressure, where getting the last stretch right needs a different kind of expertise. If that's where you find yourself, we'd like to hear from you.

What's next for Upkept

Upkept is currently handling around a hundred repairs a month through its direct-to-consumer channel, shipping to customers across 26 US states. The next stage is brand partnerships: plugging into other companies' sales and warranty platforms so that repair becomes a link a customer clicks without a second thought, rather than a separate errand.

Cash has been building the webhooks and the underlying architecture with exactly this in mind, so that Upkept can plug into a partner's systems without rebuilding from scratch each time. Robin's job, as she describes it, is finding the partners willing to pilot, iterate, and come on board without unnecessary friction.

Their ambition, in Cash's words, is refreshingly clear: "The big vision is that it's embarrassing for your company not to have a link out to repair with us." Not a nice-to-have sustainability gesture, but a standard feature that people notice the absence of.

It's a fitting note to end on for a season of HappyPorch Radio called Technology Isn't Magic. Upkept's technology matters, and building it themselves clearly mattered too. But the thing that makes the business work is the same thing that's made repair work for centuries: people who are good with their hands, given work that respects how hard it actually is.

--------

You can find Upkept at upkept.io, or connect with Robin Atkinson and Cash Acrey on LinkedIn. Listen to and read the full conversation on HappyPorch Radio, wherever you get your podcasts.