Ten thousand is the number of failed attempts Thomas Edison endured before alighting on his successful combination of gases, filaments, glass and electricity to make the light bulb.
Five thousand is the number of similar steps James Dyson climbed to invent the cyclonic vacuum cleaner. And the engine lubricant WD40 is so-called because thirty-nine failed water displacement products went before it.
Failure is at the heart of every experiment. But not just any failure. As Charles Schwab points out, there is stupid failure and there is smart failure. Smart failure contains the ability to learn. Which gets me to my favourite quote, from Samuel Beckett, one of theatre’s greatest innovators: ‘Fail, fail again. Fail better’.
No one knows anything. Today’s customers’ needs are highly global and highly flexed — and winners are constantly borrowing from each other. Experiments live in the real world and are no longer created in an ivory tower.
The value of experimenting is not simply to be first, but to be ready for when success strikes. When a competitor puzzles out a solution, you need to be familiar enough with the territory so that you don’t have to start your journey from scratch.
Learning how to experiment is a bit like learning how to grow up. Organisations are like children. Their DNA contains the combined experience of their employees, customers and heritage. But their future is unknown.
As one business leader of a pan-European chemical business told us: ‘You can’t force people to be innovative. You have to allow them to take risks; you have to allow them to fail. You have to encourage them every day.’
Like children, modern organisations need to learn through play and real-life experiences. Children start with some basic truths: heights are unsafe; loud noises are signs of danger. They explore their world, protected from self-harm by physical limits (they can’t run around for their first year) and parental limits (the permission created by their guardians). Outside of this, they are free to follow their curiosity, knowing they’ve ‘won’ by means of parental rewards (smiles), material rewards (food) or psychological rewards (sparkly things).
Experimental businesses operate in the same way. Their hypothesis takes the place of a child’s basic truth, the thing that spurs curiosity — the thing you believe is probably true. Companies that use experiments to grow are structured differently from those that don’t.
They also need to limit the harm. This might take the form of limiting investment or providing gates whereby the experiment has to deliver something in order to go through to the next stage.
Lastly, know when you’ve won. The equivalent of ‘parental’ rewards are those that fit with the organisation’s overall strategy. Google is an example of this, with an innovation culture that allows free reign but tests the results against its corporate purpose. Material rewards are hard measures such as revenue or new customers. The risk here is of imposing the constraints of everyday business.
An experiment forced to deliver on the same outcomes as existing lines of business, and which might depend on the same teams, would quickly revert to small incremental changes that deliver an upside on Key Performance Indicators (KPIs) — but ultimately fail to discover anything radically new and valuable.
The lesson here is to learn from user interface designers like those at Amazon, where requirements are set (‘Show me the “buy button” that drives greatest purchase’, for example) and experimenters are free to play within that canvas.
The psychological rewards are wholly new business lessons that were never specifically planned for. Arthur Fry had never intended to invent Post-it Notes. He had been working on a formulation for a new super-adhesive for his employer, 3M. And he failed.
Similarly, digital businesses like Facebook rarely know what to expect when they open up their data to the evolutionary forces of the developer world they just trust that great things might happen.