Failing, Fast and Slow

Failing, Fast and Slow - securely open, flexibly opinionated - Failing, Fast and Slow

(With apologies to Daniel Kahneman)

Anyone working in a large organisation must be pretty tired of hearing “fail fast” in every meeting they go to. The principle is great: you should be able to try out new ideas, technology, and ways of working, in a low-risk, low-cost manner. It’s not about failing at all, it’s about reducing the cost and time before you realise something isn’t going to work. That then lets you try more new things within the same time and budget…and you hope that something will stick.

There are lots of things you can do to help. Culture and attitude play a part, though perhaps not as much as the innovation gurus would have you believe. And anyway, failing fast shouldn’t be about changing your risk appetite, it should be about taking a portfolio approach to risk. It should also be about keeping projects small initially — we could all learn a lot from Fitzgeralds law — There are only two phases of a [large capital] program. The first is ‘It’s too early to tell.’ The second: ‘It’s too late to stop.’

So, you set up an innovation lab, or an accelerator. Or even an incubator. You can have your own teams that build interesting new stuff. You can bring in tech startups to ride around your corridors on their scooters whilst listening to 80s pop tunes. Bean bags, somewhere to park your “fixie”, table football, in-house baristas, tick! Welcome to innovation theatre — swarms of startups doing “cool stuff with NFTs”. A lot of them are going to fail, so you really need it to be fast.

The problem with innovation theatre is that you don’t know if it’s going to fail until you bring the tech through into production, so the end of each accelerator cycle is little more than an awards ceremony. Accelerators rarely get to work with actual business problems or systems, and almost never with real data. The first reason for this is data privacy and compliance — this is particularly problematic for data science startups as they never really get to validate their algorithms until they go into production. The second problem is that enterprise IT estates are invariably a mess. They’ve been cobbled together over decades and pulled in all directions as new architects, CTOs and CIOs have tried to have an influence. They’re now so complicated, it’s really expensive to integrate anything new, and adding new tech just makes it even harder to get rid of the old tech. There’s all the security and compliance as well, and start-ups have rarely had to deal with any of that stuff. All this means that it takes a very long time to pull innovation through into production, and you’re often aggregating risk all the way through — too early to tell, too late to stop.

We need a software platform that takes the friction out of innovation and pull-through. One that can bring in all our data feeds from enterprise systems, and make them available selectively and securely via modern APIs for innovation teams and start-ups to work with. A digital backbone to access the data in a controlled, secure manner so we can try, fail and who knows, maybe even succeed, with minimal fuss. Hmm… maybe the platform should also provide some enterprise must-haves such as search, knowledge graph, provenance, protective monitoring and easy deployment on-premises or in the cloud.

Luckily, we’ve got one of those. Get in touch to find out more about how our CORE platform can help you go beyond innovation theatre.