How to start a social startup: Understanding the problem

We’ve just started helping the first cohort of Bethnal Green Ventures projects and I’m using it as an excuse to write down some of the things I’ve learned about social startups over the past couple of years.

It starts with a hunch

You start the process of developing a startup with hunches about both the problem you’re trying to solve and the solution you’re going to build. In my experience these are always sparked by a story, which for School of Everything Mark I came from John Markoff’s book What the Dormouse Said,  but for other people it’s something that a friend says or something they go through themselves. The story of the Free U gave me the idea for a solution and I could quickly see the problem that it solved – or at least I thought I could.

My mistake was that this wasn’t a problem that individual people had – it was systemic. I thought the problem we were trying to fix was how rigid and out-of-date the organisation of the education system is and that is a problem, it’s just not one that a website can solve on its own. A website solves the problems of an individual person, but it then takes lots of people using the website to change the way something is organised systemically. And generally people won’t do that unless you build something that solves their individual problem.

So I think if you’re building a social start-up, the problem you’re trying to solve has two parts:

  1. an individual person’s problem that you can build some technology to help solve.
  2. a social problem that will be solved if lots of people use your solution to 1.

You can think of either one first – the important thing is that you need both.

Get out of the office

When you have these written down and your hunch about a solution, you need to get out and test them. We’ve been doing this incessantly over the summer for School of Everything Mark II. What’s important is to get accurate information. I like the analogy of this being like the scientific method: you have a hypothesis that you then test by collecting real data.

You need to think who might have the problem that you’re trying to solve. Over the summer we recruited people by asking for volunteers through Facebook and Twitter and just following our own social networks two or three degrees to get to different groups. These were as diverse as over 55s in Manchester through to young mums in London.

We found the two best tools for gathering information are Surveymonkey and shoeleather. Surveymonkey gives you some numbers but you do have to be careful in the way you design questions and interpret responses. We set the goal of getting over a hundred responses and looked for answers to be chosen by over 80% of respondents for us to think it was strong enough finding.

We also did lots of face to face interviews by getting out of the office. These give you the insights you need to know how what you’re proposing will fit into  people’s lives. You also get more accurately from these what people might be willing to pay. It’s much easier to tell whether people are serious face-to-face.

Minimum Viable Product

We also showed people what the Lean Startup crowd call a Minimum Viable Product spec. This is important because you need to be confident that you can build it. As Steve Blank says, “Any idiot can get outside the building and ask customers what they want, compile a feature list and hand it to engineering.” So as we went around asking people if anything we were reducing the number of features rather than getting more ideas about what it should include.

Once you’ve done this for a while, the answers become quite clear as to whether what you think you’re solving is a real problem and whether there are people out there who’d be willing to pay for what you’re proposing.

By the end of this process you should have:

Note you don’t have any technology yet, or a business plan or a company or full team or bank account. You can do all the above and have a pretty good idea about whether something is worth building or not but only have spent a very small amount of money.