When getting paid to build is bad

Isn’t it funny how you can be aware of a particular mistake you should definitely not make, but then still make that exact mistake, all the while convincing yourself that you are definitely, definitely not making that mistake..

That’s exactly what I did with Flask, and I got grant funding to pay me to do it! Sounds like a champagne problem? With hindsight, I think this was actually quite bad - here’s why.

The idea

Flask started from a genuine pain I had running Blue Coast. I wanted to send email marketing based on the booking take place date, not just when the booking was made. And I wanted to send different flows based on what the customer had booked.

Seem simple enough, but this actually required a Frankenstein setup of Make automations, MailerLite flows, and booking data all stitched together. It was complex, fragile, and never quite worked how I wanted it to.

I thought - I can code, I have time over winter, let me fix this problem for Blue Coast and turn it into a SaaS that can help other operators too.

The idea for Flask grew from there: integrate booking data with a clean UI for email flows, add photo management and sharing automations, and do customer analytics right, all packaged up specifically for activity providers. Basically solving my problem and some adjacent frustrations I had. My unique angle was focusing on the outdoor activity industry, leveraging my credibility from running Blue Coast.

The grant

Local grant funding came up for digital development projects. I thought why not get paid to develop this for Blue Coast? Can’t go wrong, perfect timing.

I had found out about the grant late and had a tight deadline, so I scrambled to scope out what I’d build and submitted my application. I was successful, and pretty stoked to get stuck in building the product.

I spent most of the winter of 2024/2025 doing just that, and because I had been grant-funded to produce what I’d applied for, I felt obliged to build to that scope.

Months spent building a reasonably complete application based almost entirely on my own desires. Not a discovery interview in sight.

“Evidence”

I did look for validation, but I was looking in the wrong places.

My market research consisted of posts I’d seen in Facebook industry groups for tour operators. There were other competitors building similar solutions. Surely that meant there was demand?

My assumptions stacked up: there’s demand in the activity industry, tour operator needs are similar to activity provider needs, and “I am my target customer”. I was getting paid anyway, so I just cracked on and didn’t worry about it too much.

The problems

Let’s count the flawed assumptions.

Tour operators are not outdoor activity centres. This was already quite a niche idea that only a segment of tour operators might want. For actual outdoor activity centres? I hadn’t tested that assumption at all beyond “I want it”.

But I am not my target customer. I do own an outdoor activity centre, but I’m also technical, interested in marketing, and willing to build and maintain complex automations to solve problems.

Is there a large enough segment of outdoor activity operators who want to set up automated personalised marketing? Kind of feels like an oxymoron, and with hindsight I think it was an idea doomed to fail from the start. But it was pretty easy to convince myself otherwise, especially with that sweet grant money coming in.

The real cost

Getting paid is nice, and if the goal was simply to get paid for two or three months of development, then it would have been a happily ever after.

But the goal was always to build a viable business that stands the test of time. I spent months doing not that while convincing myself I was - the real cost is the opportunity cost. (It’s painful to think about how much further I’d be right now if I’d spent that time differently.)

It was only after I’d built out a fairly complete MVP that I started talking to potential users. And then I realised there wasn’t a market for Flask in its initial form, or at least not a market big enough to make it worth the effort.

Some other problems surfaced as well, but critically I more or less put the blinkers on and ignored everything else while steaming full speed ahead developing based on untested assumptions and an unvalidated idea.

Lessons

• Grant funding can be dangerous if you’re applying too early. Grants that require scoped deliverables can force you into building before validating. You need to validate your business model and speak to actual customers.
• You are not your target customer. Even if you really think you are. There’s something different about you, you’ve decided to build software to solve your own problems. That doesn’t disqualify you from building for your industry, but it definitely makes it a lot easier to convince yourself that because you want something, the market needs it.

Pivoting

Flask has now pivoted into Swellbase.

I’ve done things properly this time (and I’m sure have made some mistakes I’ll benefit from some hindsight on later!). 40+ centre owners have been kind enough to have a chat with me on the phone over the last few months and I built a much more extensive business case before writing the first line of code.

Some things from Flask were reusable, some lessons carried over, and some problems were solved for Blue Coast in 2025. It’s easy with hindsight to look back and think I would have done things differently. But I probably couldn’t do Swellbase the way I’m doing it now without going through the Flask experience first.

Maybe you can skip the detour.