Unicorn Farming: Building Capabilities Against the Odds (Part I)

When data leaders meet at conferences, organisation structures are a key topic of discussion. Who do you report to? What teams do you have? And of course: how is your team resourced? This last question is usually met with an uncomfortable laugh followed by, “Our team’s headcount? With or without all the open roles?” Rarely do you meet a chief data officer or data scientist who is not actively recruiting at that moment. We all are, all the time. As McKinsey found, the number of organisations developing AI solutions is accelerating.

Talking about recruitment is to data leaders what talking about the weather is to the British.

We know the reasons: big data is big business; we are in the middle of an AI arms race; data visualisation is the new black. Our definition of analysts is changing: they no longer wield Excel and pivot tables, but instead wrangle Python and pandas. The world is speeding up and with it our ability to keep pace from a capability perspective is decreasing. To close this gap we need unicorns: people gifted with technical skills, business acumen, and a transformation mind set. Sadly, finding unicorns is hard and if you do not work for a hot start-up, the odds are against you recruiting and retaining one.

There are many stories we tell about how we ended up in this predicament. However, I have come to feel that while discussion about talent shortages make for good panel fodder, the reality is not quite as bleak.

Firstly, as much as new trends are reshaping industries, they move slower than people might think. For example, it took years for mobile browsing to supplant desktops, just as voice will take years to become a meaningful mode of interaction for consumers. Yes, the potential is huge. But right now most people only use their Alexa or Google Home to set timers and play music. This is not to say changes are not happening, but that the speed of change is slow enough for companies to adapt — if they focus on building capabilities.

Secondly, the specific technologies matter less than people tend to believe. While message boards are filled with discussions — on what programming language to choose, which cloud platform to leverage, or even what courses analysts should take at University — it mostly comes out in the wash. Given how commoditised cloud offerings have become, you can build data products and perform analyses whether you use Python, Java, C#, Go, even PHP. Additionally, once you know one programming language it becomes significantly easier to learn another on the job — that includes knowing VBA!

What does this mean for organisations looking to build analytics and AI capabilities? As the cover of the Hitchhiker’s Guide to the Galaxy reads: “Don’t panic.” While recruitment might seem like a daunting prospect, the worst thing one could do is give in to procrastination or paralysis. Know that building a data capability and attracting the right talent is possible, whatever your business. There is enough time, if you make a concerted effort and avoid a number of talent related pitfalls. We will review those in next week’s article.

— Ryan