What I learned doing a startup: part 1

This post comes from a series; read the series overview here.

Validate your validations

When I began Ground Control I had no idea I was starting a business. I had been hired to build something, and wanted to do a good job of building it. However once other people expressed interest in the same system, and realising my initial customer would allow the software to be licensed in this way, I got the idea that this could become a system that could be sold over and over again.

The first customers paid lump sum values to use the system, but I quickly realised that large one-off payments for software had disadvantages - they made the sales process harder (the customer is committing a lot up front) and they risk the business needing to keep selling to survive. This would be a problem for any business but skydiving is a small industry - one could feasibly run out of customers, or at least exhaust all potential leads.

The next step after moving to a monthly subscription model was to set a price, and start figuring out who’d pay it - did I actually have a business model? So I needed to validate my business against the market to understand who’d buy it and for what price. In doing so I made a key mistake: I thought I was validating my market, but never checked what I was validating against.

Validate your businesses

The easiest way to validate a business is to build the business and then go and sell it. The problem with this mechanism is that you need something to build the business with - either time or money. Fortunately by the time I was creating a business I actually had a product, so I could take the easy route.

My thought process was relatively simple: choose a price, then ask people to pay it. If they didn’t like it, lower the price until they did, or until it seemed silly to go lower. I set my first price arbitrarily and this was a mistake - I guessed with no evidence at how many customers I could reach in a few years, and set a price such that if I reached that many customers I’d be making enough money to live off. What did I neglect?

  • Business growth tends to require growth of customer base and growth of headcount. The two don’t need to scale in the same multiples, but there’s a feasible ceiling of customers one can service as the sole employee. If your pricing matches your desired wage, there’s not room for hiring
  • Business software isn’t art. You’re not trying to extract the time value you put in, you’re selling the time value the customer gets out of it. If your software saves a staff member an hour per week it’s worth their hourly wage x4 per month, or thereabouts. Not all software is as easy to measure, but broadly you’ll give business value and it needs to be equal to or greater than what you charge
    • An exception to this may be in a form of “gatekeeper” business - your product is so essential that there’s no value proposition, customers have to buy it and you just need to buffer versus your competition and what customers can feasibly afford.

Nevertheless I proceeded; my validation required that I offer the subscription price and people accepted. For 2 years this happened, some times to lengths I didn’t think possible - whilst at one airfield they got a call from another in a different state, and one phone call later I was on a plane to visit them, gaining two more customers on the trip.

Then I hit a wall; a 1,300 mile road trip round California and Florida visiting an array of small and large airfields, and nothing to show for it.

Validate your validation

My validation model was simple and flawed. If people will buy, people will keep buying. It’s not quite “build it and they will come” but more like assuming the visitors wandering up the path are endless, the line snaking through the next county. Because of how I’d marketed (word of mouth mainly) I’d reached people who were similar - who saw the value of software, who saw the benefit of a hosted system when other vendors were offering on-premises solutions, who saw value in a system built for the industry rather than generic booking or accounting tools.

Those people were great and they were great customers, but I either ran out of ways to reach more people like them, or I ran out of people in that mindset. During various sales visits I was asked to do the system for £10 per year, was asked multiple times to sell a lifetime license, was asked to hand over the code to run on-premises, or asked for things I couldn’t yet provide, like 24/7 support (I basically did provide this because I cared about the business, but I wasn’t silly enough to put it in writing).

I realised that I hadn’t thought critically about my key assumptions, against which I validated my business:

  • I was charging low hundreds per month. This worked out as a straight value swap with estimated staff time, but the businesses didn’t work like that - they often had a mix of long serving and very temporary staff, and often employed friends or family members. So offering software as a way to cut staff costs didn’t fly.
  • Most skydiving centres do not operate a particularly happy cash flow - there a boom and bust times every year or even month-to-month. Even if it can smooth business in the big times, a monthly subscription just looked like a cashflow risk to many.
  • The system was comprehensive and thus valuable, but it was hard to get that all across clearly enough before the entry price was asked about. That tended to stop conversations when people are used to “things on the internet are free”.
  • Software subscription models were a relatively new thing. When I decided on this approach I was reading TechCrunch, attending tech networking events by companies with cool logos and VC investment. My customer base was not these people; they were intelligent but happy with what they already knew of how they bought software.

So it turned out my model wasn’t valid and it was too late to do much about it.