Data Science in the real world
Data science. What does it actually mean? My personal view is that it should be used to describe the more advanced analytical techniques that we now have in the industry arising from the increased availability of data (and also the infrastructure and tools available to process this).
Think machine learning, A.I., and so on. Equally though, the term ‘data science’ seems to be rather lazily used as a description of the more standard analytical approaches such as regression or clustering (I’ve even seen it used to describe the creation of Excel pivot tables).
Anyway, irrespective of what it covers and who’s using it, one thing remains constant. There is no benefit in data science, modelling, machine learning, regression, K-Means, or clustering unless it delivers commercial benefit for an organisation.
On this point, it seems that whilst the data industry is finding new and interesting ways to interrogate and analyse data, our ability to translate this data into insight – making the complex simple – is not developing at the same rate.
This is a unique skill – I’m talking about the person that can describe what a neural network does to a CMO in words of less than two syllables.
It’s not necessarily a skill that you learn through academia. It needs experience of the commercial world, along with experience of the analytical techniques themselves. It’s not necessarily a skill that a Data Scientist would possess either (other than the very best).
It’s a tough ask to get a Data Scientist to be focussed on delivering the best possible analytical solution but then ask them to step out of that analytical bubble into the shoes of their end client.
I’m not talking about Planners here either. It’s typically a step before a Planner starts to really get involved in a project, kind of a bridge between the analytics and planning.
So exactly what skills does this insight role need and how, as an industry, do we develop more of them?
The start point is a solid grounding in analytics, either academically or commercially (preferably both). More often than not, these individuals have been hands on analysts themselves and so they appreciate and understand the challenges a data scientist will come across when delivering against a client brief.
Coupled with that, a planning mentality is essential – a real desire to get under the skin of a client’s business, how it works, how it makes money and what its challenges are. This is essential to ensure that the analytics delivered can really meet the brief.
Data and technology are becoming much more closely intertwined in the data science world and as a result, these people need to really understand different technologies and the benefits they deliver.
The growth in open source technology means you simply can’t get away with just having a pure SAS or SQL background any more – you need to understand the tools your data scientists are using and their relative merits.
Perhaps the most overlooked of all their skills is storytelling – translating outputs and findings from a project into real commercial gold. I recently interviewed a candidate who had a first-class maths degree but also grade A A-Level English. I know people like this are few and far between but that’s almost the perfect background – numerate and literate.
Finally, these individuals need to feel comfortable in front of their stakeholders. There’s no point creating a commercially valuable piece of work if you can’t land it with your client. Again, this is a skill that’s not always prevalent in the introverted world of analytics.
So that’s it – I’m not asking for much am I? If you can get people like this in your organisation to work alongside your data scientists, you’re onto a real winner.
We’ll use your information for the purpose of our newsletter, to occasionally invite you to our events, and send you other relevant information..