data science
Futuristic infographic with Visual data complexity

This month, David Socha discusses the true value of Data Science and the opportunities being missed, as businesses stall at the very point that insights turn into value.

In recent times, I’ve met with a lot of businesses that tell me they’ve got Data Science covered and that I can probably save myself some time and effort by not bothering to discuss it any further with them.

From around 2013-16 those same businesses would tell me that they’d got Big Data covered. Because they’d “got some Hadoop”. Before 2013-ish, they certainly had Data Warehousing covered, because they’d bought SAP Business Warehouse. Oh, and Business Intelligence and Analytics were also all good. Still are too, thanks. Did we even mention Cloud? We’ve got some of that as well. It’s great. The best Cloud. You’d love it. Seriously.

Sounds like I’ve had lot of brush-offs huh? Maybe some of you are even starting to think you’ve been unwittingly drawn into one of those ...eh… inspirational essays that fill up your LinkedIn feed about focusing on the positive learning experiences of being fired and having to give up your beautiful German car and holidays in the Caribbean and how the kids have benefited so much by leaving behind that fancy private education to go to the School of Hard Knocks. But don’t worry, that’s not where I’m going with this, I promise.1

Here comes the science bit

Let’s get back to Data Science. And let’s revisit those conversations with businesses that are already doing Data Warehousing and already doing Big Data and most recently already doing Data Science and Cloud and such too.

Typically, what does that really mean? In 2017-8 it often means that a few of the younger members of the organisation have access to some “free” tools like Spark, Python, R and maybe TensorFlow and have persuaded you to put some data in the Cloud.

Because paying for the amount of data they’re going to need to be on Amazon Web Services is such small change in the Departmental budget that the cleaners can probably sign it off.

And you should see what they’re coming up with!2 It’s amazing! We can predict when an asset is going to fail! We can see clear paths to customers not paying their bills! We’re changing the world! And it’s free! OK, yes, it is using very small sample data sets and it is only running on a public AWS instance and you can only access it via young Georgia’s laptop. But we’re doing Data Science! The Chief Exec is excited!

Now, I don’t mean to be rude. All of this should be happening. This kind of experimentation and discovery is precisely how data and analytics can show its potential to underpin business improvement, business change, even new business models.

But here’s the thing: making real change requires much, much more than downloading a few tools and giving some data to an intern with a laptop. Or even setting up a whole new Digital Division, if its only purpose is to have more young people with cooler laptops than the rest of us so the CxO can tell his friends he runs a cutting-edge business. Delivering any kind of real value means operationalising your insights.

Scale: the Holy Grail

A colleague of mine recently wrote a blog that at first glance is about AI. In reality it’s about finishing what you’ve started. I echo his sentiments here and when I’m in those conversations with businesses that tell me they’ve got Data Science covered. Of AI, he states:

“That next wave can indeed be a game-changer, but it will just be a ripple unless the analytic platform is built to function at scale.”

Swap out AI for Data Science. Still make sense? Of course it does. Data Science can be a game-changer. It’s not a transient technology, like Hadoop may yet prove to be. And it’s not just a trend. It’s a way of thinking about data and business problems and addressing them with the tools of the day, whatever they might be now and in 2020, 2030…whenever.

But Data Science can only be that game-changer when insights are applied. Applied to full data sets; applied to actual business operations; applied to things that are measured and that can be improved. Applied at scale.

Scaling up is a very different problem of course. It involves infrastructure. And planning. And commitment. It involves budgets, too. But there’s no getting around the fact that it’s where the value lies.

Next month, I’ll talk about how to achieve that transition from interesting discoveries in the Data Science Lab to moving the needle on real business KPIs. Until then, next time someone comes to talk to you about Data Science or Advanced Analytics, or AI or indeed any other method of getting the most from your data, think carefully about what you’re really doing in that space before you show them the door.

Are you really delivering on the promise of the things you’re trying out today? Are your insights changing how you do business? Or how you do anything with a KPI? If not, is there at least a roadmap to get there? A funded roadmap? Because if not, you’ve probably been wasting your money. See you next month.

1 Having said that, I admit that some of these essays probably do have a point at times.

2 The Data Scientists, not the cleaners.