When you’re starting out with something new, it’s important to understand that your customer is never “everyone.” Even if your eventual potential market is huge, you need to start out by dominating a small niche.

Google didn’t start out by organising the world’s information.  Google started out as a way to make searching the Stanford Library easier as part of the Stanford Digital Library Project.

Facebook didn’t start out aiming to connect everyone in the world with cat pictures and artificially manufactured political outrage.  It started out as a way for Harvard students to hook up.

I ran into more examples today in Carlota Perez’s brilliant book Technological Revolutions and Financial Capital. Perez outlines the dynamics of technological and economic change across five revolutions, and she talks about how new technologies start out by fitting into small parts of the existing economy:

…they grow restricted to whatever uses fit well in the existing fabric of the economy before their most important uses are even surmised. Railways were first developed to help get coal out of mines; their real significance as the main means of transport of people and goods was difficult to even imagine in a world of canals, turnpikes and horses. Oil refining and the internal combustion engine developed within the steam-engine world of the third revolution, being used mainly for luxury automobiles. Semiconductors, in the form of transistors, served to stretch the market for radios and other basic appliances of the mass-production paradigm by making the portable, before anyone could possibly conceive of a micro-computer.

We’re currently running a small Lean LaunchPad course at University of Queensland, and listening to the updates on Friday got me thinking about this all over again – all the teams were aiming too broadly.

This reflects several important issues.  The first is that they are all working on big problems, with big potential impact – so it is natural to aim for the biggest possible market right from the start. Unfortunately, this approach fails.  We need to work on big ideas, with big potential impact, but we have to start out in the smallest possible application.

The reason for this is that we don’t know in advance how to make genuinely new ideas work.  This is one of the reasons that the flat part of the innovation diffusion curve is flat – it takes time to work out the best value from a new idea, and the best business model to use to realise that value.

Making a new idea work requires three distinct sets of skills.  First, you need to have the skills of creativity and invention to get the new idea to work in the first place during the invention phase.  Then, you need to use your customer development and problem-solving skills to create a market with the early adopters.  Finally, you need different skills again to make the transition to a business model that will scale with mainstream customers. This is the problem that Geoffrey Moore called Crossing the Chasm – illustrated nicely in this post by Peter Armstrong:

The three sets of skills don’t always live within one organisation.  In fact, they rarely do.  This is why we often see different companies dominate at different points in the industry lifecycle.  The automobile was first invented by many different people, none of whom are remembered today. Then, as Perez points out, the early adopters were wealthy people, and the first company to figure out how to successfully serve that market was Duryea Motor Wagon Company in the early 1890s.  Fifteen years later, Henry Ford figured out how to make cheap cars that the majority of people could afford.

We’re seeing the same thing right now with autonomous vehicles.  Someone (who?) invented the idea, and Google (or Uber, or someone else) will end up taking them into the majority.  But they’re already in use right now in mines.  This is where the need is currently most acute, and this is where the technology is being refined.

Eventually, all new buildings will be constructed using Building Information Modelling (BIM) and the offsite manufacturing of modular components.  But right now, those technologies are only being used in the most challenging settings for construction, like the Leadenhall Building in London.

Autonomous vehicles and BIM will eventually both be huge.  But today, they are really small – and that’s the only way to eventually win.