If you are thinking of adopting data science into your organisation - and you really should be - then you are going to need to create a data science lab. The reason? Your existing processes are not fit for this purpose and the people you will be employing in the lab would not want to work within them.
If much about existing business culture has proven resistant to conventional data and analytics, then the reality is doubly-challenging when it comes to stretching the boundaries even further. Decision makers and data “owners” within organisations often prove surprisingly resistant to the notion of having an evidence base, pooling their information to gain uplift everywhere and, most particularly, placing their actions under the cold, hard scrutiny of proper metrics. Politics and human nature get in the way of reason.
When it comes to data science, that is going to be amplified even further. One of the key dimensions of this practice is to look for any data that could help to build a more effective model, both inside and outside the enterprise. Weather, bank holidays, news events - all of these could end up being highly predictive and outperforming existing benchmarks which are based solely on internal indicators. That can be an uncomfortable experience for managers and senior executives whose judgement and historical decisions get called into question.
As a result, projects that want to explore the potential benefits and impact of data science risk being cut down if they are developed within the existing business parameters. Enter a new generation of labs with a completely different culture - not just digital in style, but hacker in ethos, rapid and flexible in approach, and challenging even to the project leaders who commission them. Put that in front of a board used to horizon lines for transformations which are measured in years, rather than weeks and you can understand why there is a problem.
To get around this, a new generation of start-up consultancies is evolving which puts data science into the category of disruptive advisors. If your board is willing to listen to a tax advisor presenting a leading-edge way of minimising the HMRC bill, then it already understands how to hear advice that is outside of its comfort zone (and often unproven until undertaken). Even more so if it has employed a consultancy to advise on digital transformation. The outputs from data science labs fit that model, since they can involve radical re-alignment of resources, rethinking business processes or targeting entirely new market segments.
Practitioners do not sit comfortably alongside your existing data and analytics teams. They probably don’t even work that well within Shoreditch digital-style environments, however much these latter practitioners believe themselves to be the cutting edge of cool, disruption and creative innovation. Only by challenging for a space currently inhabited by the big league management consultancies can data science labs get into conversations at the right level. What they claim to offer is not only as big, it is also often easier to operationalise. Ever tried to turn a consultancy’s playbook into live business processes?
That is a big play, yet it is one that a growing number of these highly-educated and innovation-minded individuals are willing to make. Increasingly, they are finding backers to fund them through the difficult first few years. If you are not already planning to employ one of these data science labs - or to build one as a new capability somewhere in your enterprise - then all you can do is sit back and watch somebdoy else rake in a big pile of chips.