Alphabet (the business formerly known as Google) has become infamous for its project culls. Shiny programmes that look from outside like enviably serious businesses, either stripped back or canned altogether. Its drone programme, Project Wing, seems to be the latest (Bloomberg via Benedict Evans’s newsletter). Evans suggests that following a spread of bets across potential game-changers, Google knows now that machine learning is the future and so it can double down on that*.
But that doesn’t mean that the other bets Google made, and has pulled back from, couldn’t be serious businesses in their own right. Someone is going to make a lot of money selling (or perhaps leasing and operating) delivery drones. As the Motorola Moto Z I’m currently testing ably demonstrates, modular smartphones may be viable, despite the death of Project Ara.
Google may have selected the best opportunities. But the others are still opportunities.
Second Place is Great
I’m doing a little piece on local radio this morning about the death — or otherwise of the DVD. A well-loved film store in Kent is apparently closing because of lack of demand. It’s no great surprise: sales of video discs have halved in the UK since 2008 according to IHS. That’s a tough market in which to sustain a small business when you’re likely being undercut by the nearest supermarkets, and outsold by the internet.
But I’ll be arguing that the DVD won’t go away. Not for a long time at least. It will be diminished: digital streaming will take the lion’s share of film sales. But there will be a hard core clinging on, for whatever reason. Like the US prisons, highlighted recently by 99% Invisible, that still use cassette tapes.
In our globally connected, diverse market, even the smallest opportunities can be significant.
This or That? Both
This reality doesn’t offer the sort of certainties people like.
“Is the answer A or B?”
“Well it’s about 40% A, 25% B and have you considered C and D?”
But in all spheres of our life we are going to be experiencing, and taking, more choices.
That includes our working environments, something I’ve been looking at again recently due to some engagements in the property sector. Thanks to the Workfit-T from Ergotron I’m currently testing for The Loadout, I now spend some of my day standing, and some of it sat. Like most people my days are also divided between working in the office (less) and working remotely (more). A mixture of my kitchen island, coffee shops, conference centres, hotels, railways, airports and other people’s offices. There are not going to be office workers and remote workers — especially since the office-bound jobs are the most likely to be automated. We will all be on the move.
The best strategy in this diverse and complex environment is to embrace as many good options as your resources allow. This doesn’t mean chasing every avenue. It means selecting the best opportunities early, experimenting, and learning to cull when you’ve learned what you need to learn. It means adapting when new opportunities you hadn’t considered become available, even if the idea wasn’t yours. And it means learning to be comfortable with complexity and some level of disorder. This is not easy for us as a species: we like to put things in boxes and draw dividing lines: this OR that, not this, that and the other.
In business we can only afford to support this complexity if the interfaces between us and our many choices are low-friction. Small opportunities can be made unviable if the cost of addressing them is too high.
Inside a company this is often about measurement and decision-making: so much energy is expended in making decisions that there is too little left for the practical experiments that might inform them.
At the borders of the business, friction is often the difference between being able to pursue lots of opportunities and having to make a few big bets.
Lower your friction and your business can be much more secure.
*This would be consistent with Google’s continued investment in human computer interaction technologies that can increase the bandwidth of the input and output of data from these learning systems.