In recent years, a lot of traditional companies founded digital labs that should server their digitalisation efforts. But if you look at the results of these labs, they are rather limited. Most of the “products” or PoCs never came back to their real products. Overall, they could be considered as failure. But why? Why is it the wrong data strategy?
What is a silicon valley lab?
Let’s first look at what those labs are and why they were founded. Everywhere in the world, there is increased pressure on companies to digitalise themselves. Basically, how C-Level executives handed that in traditional companies is by looking at (successful) Silicon Valley startups. They did trips to the Valley and found a very cool culture there.
Basically, a lot of companies were built in the garage or old fabrics, thus giving that a very industrial style. Back in good old Europe, the executives decided: “We need to have something very similar”. What they did is: they rented a fabric hall somewhere, equipped it with IT and hired the smartes people available on the markets to create their now digital products. Their idea was also to keep them away (physically) from their traditional company premises in order to build something new and don’t look too much at the company itself. A lot of money was burned with this approach. What C-Level executives weren’t told are a few things:
(A) Silicon Valley companies don’t work in garages or fabric halls because it is fancy and the way they like it. Often, they don’t have money to rent expensive office space, especially when prices in the valley are very high. The culture that is typical for the valley is rather something that was done because it was necessary, but not because of coolness.
(B) being remote to the traditional business works best when you develop a product from the ground up and completely new. However, most traditional companies still earn most of their money with their business and in most cases it will also stay like that. A car manufacturer will earn most money with cars, digital products then come on top. With this remote type of development, it often proved impossible to integrate the results of the labs into the real products.
So what can executives do to overcome this dilema with failed PoC’s in Data Science projects?
There is no silver bullet available for this challenge. The popular website Venturebeat even claims that 87% of Data Science projects never make it into production. It depends mainly on what should be achieved. When we look at startups, their founders often come from large enterprises that were unhappy with how their business used to work.
I would argue that most large enterprises basically have the innovation power they are seeking for, but it is often under-utilized or even not utilized. One thing is crucial: keep the right balance between distance or closeness to the legacy products. It is necessary to understand and built on top of the legacy products, but also it is necessary to not get corrupted by them – often, people keep on doing their things for years and simply don’t question it.
To successfully change products and services, the best thing is to bring in someone external that doesn’t understand the company that well but has the competence to accept their history. This person(s) should not be engineers (they are also needed) but rather senior executives with a strong background in digital technologies. Seeing things different brings new ideas and can bring each company forward in the digitalisation aspects 🙂
This post is part of the “Big Data for Business” tutorial. In this tutorial, I explain various aspects of handling data right within a company.
Leave a ReplyWant to join the discussion?
Feel free to contribute!