If you believe the headlines, data is the new oil. The best job rankings have been oscillating between data scientist, machine learning engineer, AI specialist, or something similar for the past decade. It’s Data Wonderland. And just like in Lewis Carroll’s Wonderland, if you don’t know where you’re going, any road will take you there.
I mean this as satire of course, but if it wasn’t immediately obvious do go ahead and read the original exchange between Alice and the Cheshire Cat so you can appreciate its genius before I explain my own joke (and ruin it). If you already know where this is going, feel free to skip the quote, and the paragraph after that.
“Would you tell me, please, which way I ought to go from here?”
“That depends a good deal on where you want to get to,” said the Cat.
“I don’t much care where — “ said Alice.
“Then it doesn’t matter which way you go,” said the Cat.
“ — so long as I get SOMEWHERE,” Alice added as an explanation.
“Oh, you’re sure to do that,” said the Cat, “if you only walk long enough.”
The punch line is that if you don’t know what you want to obtain from your data science efforts, or (worse yet!) don’t care much where you get to, chances are you’re not going to be content with your destination.
So, how do you ensure your data science team knows where they’re going? Ask yourself the following questions: Do you have a data strategy? Is everyone in your organization aware of your data strategy? Are the key players in your organization on board with your data strategy?
If you’re not sure on either of the 3 points above, you need to do some strategy work before you can expect to get value out of data. For a data strategy to be robust, I have found three aspects to be key:
- PURPOSE — this is what drives you. Why are you doing this? If the answer is along the lines of “everyone else is doing it” (typically your competition) that is a good observation, but it does not amount to purpose. One way or another, your purpose must be linked to your core business strategy if it’s to be of any use at all. And what is most important is that it’s clearly articulated. Do you want to increase customer satisfaction, make your operations more efficient, retain key employees? Those are all good and frequent example statements of purpose for analytics and data science. You just want to get started with some ‘data mining’ and see what comes out of it? Definitely not a good statement of purpose, and it could be setting you and your team up for failure.
- MEASUREMENT — are you measuring what you hope to manage, and are you measuring it well, leveraging all the tools at your disposal? Your purpose is very much necessary, but it is not sufficient. Up until your purpose has been followed up with a sound plan for what you are going to measure, how you are going to measure it, and how you are going to ensure the quality of your measurement, and up until that plan has been put into action, you merely have good intentions. As they say, the road to hell is paved with them.
- EXPERTISE — do you have the right people to analyze your data in a way that does justice to your purpose? Have you involved these people from day one? For many organizations, it becomes apparent only when they bring in the experts that their purpose is not well-articulated and the measurements they’ve been taking do not really enable them to size up a problem or offer a solution. Don’t just call in the statistician to conduct an autopsy on your data! Alternatively, if this advice came too late, do read my article on “bad data” and how to get it to work for rather than against you.
Data strategy takes input from purpose, measurement, and expertise, then binds it all together in such a way that the end result becomes more than strictly the sum of its parts.
For practical reasons, a lot of data initiatives separate purpose, measurement, and expertise, not just conceptually, but also at the team level. When doing so, there is a temptation to overlook the data strategy role, reducing it to a hat that different members of the team sometimes wear, according to their own preference and experience (or lack thereof). This reduction of the data strategy role is likely how your data science team ended up in Wonderland. Coupled with a strong top-down culture, this can become a particularly dangerous game of sacrificing effectiveness for the sake of efficiency.