What I learned doing a startup

I’ve considered writing up some thoughts about my first business for some time now. There’s become a tradition in the startup world of publishing post-mortems when a company fails. In many cases these are published after the startup may only have been around for months, or a low number of years; they’re usually published once the business is wound up or does a dramatic pivot which essentially amounts to it becoming a new business. My case is somewhat different; Ground Control is a business I started in 2010 and exists to this day, but it shared a similar aspect as other “failed” startups in that it never got to the place I imagined it could be when I started.

So I want to write about that for two reasons:

  • Every startup journey is different; challenges can be broadly applied but will be unique. This means having had the experience there is likely to be something in there that can appeal to or even help somebody else in a situation which, if not similar, is definitely parallel.
  • A lot of what I write or talk about is learned knowledge. It’s stuff I’ve read up on, watched videos on or worked on directly and am trying to simply re-distil for whatever audience I am able to reach. In the case of my business however, I’m also wanting to ensure I actually know what I learned - and a good way to do that is to write about it.

To keep things simple I plan to write one post per lesson; I’ll use this article as both an introduction and an index. There’s not likely to be a specific order to the posts, though given I have to write them in some order I’ll number them anyway, and its possible I’ll assume you’ve read post N before N+1 and so might reference things from previous posts.


  1. Validate your validations
  2. The customer isn't always king
  3. Early decisions do matter

Check back for more posts in the near future…


What was my business?

In 2010 I graduated from university and with the UK still on its path out of a recession decided to try my hand at freelancing rather than trying to get a job. I had built a few websites on the side whilst at university, and as such had a business registered, bank account set up and some portfolio work to sell my services. I had a few interesting experiences trying to pitch websites to SMEs in the Manchester area but the work never amounted to much. As part of that, right near the start, I pitched building new customer-facing websites to every UK skydiving centre (see here for more of my skydiving exploits). Out of >20 businesses emailed, one came back to me.

The business that responded wasn’t interested in what I was offering, a new website, but rather an off-hand comment I’d made at the end of the video: “I could also write a system to automate the day to day work at your dropzone”. This idea had been with me for a while, since seeing the cranky old desktop software that was basically a front end for Microsoft Access that the dropzone I had trained at used. Back in 2010 the “software as a service” business had been in place for a while but wasn’t the all encompassing world it is now - nevertheless I had a big conviction that websites that accomplished business tasks were a better way to go than just more marketing, e-commerce or social platforms. One skydiving centre thought the same and Ground Control was born.

The initial build took about 3 months; I wasted a lot of time trying to do fancy graphical designs before nailing the core functions, and gave very little thought to the code architecture. My first demo to the customer was a big let down, but one advantage of being young and single was that time wasn’t an issue. I spent weeks at a time living on an airfield, watching people use the software during the day and doing a sprint-worth of changes overnight. Within 6 months the system was running day-to-day operations, and another centre had got word and wanted to start using it.

The next centre had an additional requirement that the system would handle all their bookings made over the phone, in person and via their website. I set to work making a lot of changes, and in the next 3 months further centres got wind and I began to deploy the system more widely. By the middle of 2012 4 centres were using the system full time with a 5th in progress; then I met a USA-based skydiving centre owner who wanted the software state-side. I decided to wind up my other freelance activity, added an option to render dates the wrong way round (m/d/y seriously?) and got on a plane.

That was a lot of words, anyway, what went wrong?

By early 2014 Ground Control was in use at or being trialled by skydiving centres in the UK, USA, New Zealand and, for 2 weeks per year, Panama. I had a range of different contracts and pricing which was confusing and difficult but I was making money - just enough to live off at least. However I needed larger clients and probably some help if I was to make more of it. In March 2014 I went on a 1300 mile road trip around both California and Florida visiting centres to pitch the software.

I met a lot of interesting people, had a lot of positive conversations and checked in on a few existing clients. At the end of the trip I had a promise by a large centre in Florida to check me out. I spent 3 months setting things up and training, and it all went wrong. I can detail why in follow-up articles, but the experience convinced me that the way I was doing things and the business I’d built didn’t fit in the industry I was trying to sell it to. I explored some possible pivots but was running out of money. Shortly afterwards I was offered the CTO role that I currently have, and it made the most sense to accept.

Where’s Ground Control now?

Ground Control still exists - any skydiving centre can set up a free account. The servers run, the data gets backed up, the code has relatively few bugs, and most centres using it have been doing so for over 5 years. I never planned to shutter the system but always had the hopes for some resurgence, however the lack of ability to make money coupled with growing technical debt means large changes are unlikely at this stage.

I continue to communicate with and support my customers in the limited time available to me, and when I visit skydiving centres I still talk about what I did/do and new centres still take advantage of what the system offers.

So Ground Control is far from a failed business, but it’s equally hard to label it a success. In some ways it feels like a prototype that made it to production (which I’ve learned is pretty damn common) or just an idea that was a bit too big for where I was when I started.

What I know now is that if I started again, I’d go about a lot of things differently. That tells me I’ve learned things in the process. I hope you join me to figure out exactly what.