I admit it: I have a conflicted relationship with data.
On the plus side, I believe that data is one of the three biggest benefits that the digitization of the world has given us marketers. (The other two are malleability, the ability to easily change, experiment with and personalize the digital canvas, and intimacy, more direct interactions with customers.)
I believe that good, data-driven decision-making can compensate for many of our irrational mental biases and lead us to better outcomes.
And, I believe that every modern marketer should become adept at using data properly. If you’re not harnessing data to make your marketing programs better, you’re missing the 21st century.
Data Fundamentals Vs. Data Fundamentalism
However, I don’t believe that data has all the answers. In particular, I get itchy when I hear people imply that data will generate their strategy or creative.
Analyzing the linguistic data of Harry Potter and The Cuckoo’s Calling to discover, with high probability, that Robert Galbraith was actually a pseudonym of J.K. Rowling is a great example of what data and machine learning can do. Actually writing Harry Potter or The Cuckoo’s Calling is a good example of what they cannot do.
I believe that data can provide the inspiration for new ideas in strategy and creative, especially when we take an exploratory approach to all the data we have available to us. And, I believe that data, especially in the context of a controlled experiment, can provide validation for those new ideas.
But, I do not believe that strategy and creative will predictably emerge from the data of their own accord. On the supposed end of theory that big data will bring us, I call, um, malarkey.
Kate Crawford wrote a blog post on the Harvard Business Review site last year, The Hidden Biases in Big Data, where she labeled this unrealistic faith in data to have all the answers as “data fundamentalism.”
I believe in data fundamentals — using data in the right way, in the right context. But I eschew data fundamentalism. Knowing what data cannot do is equally as important as knowing what it can.
The Map & The Territory
I also get itchy when people claim that data is truth. How many times have you been in a meeting where someone brandishes a statistic to “prove” their point of view — without weighing the caveats and context in which that statistic was calculated? Things like selection bias, sample size, confidence intervals, and confounding variables matter.
Statistics show that people love content marketing. But do the characteristics of the content matter? Do the characteristics of the audience matter? (Hint: yes and yes.)
An often repeated phrase, attributed to Polish-American scientist and philosopher Alfred Korzybski, comes immediately to mind: “The map is not the territory.”
Just because someone drew a map of a mountain, doesn’t mean that is how the mountain actually is in reality. The map is obviously a simplification — otherwise the map would be as big as the mountain itself. So, there inherently must be many things missing from it. Things may have changed since the map was drawn, say, due to an avalanche or an earthquake. Or the cartographer may simply have made errors — in observation, in translation, in judgment or in drawing the map itself.
The practical wisdom of Korzybski’s insight: if you come to a cliff, don’t keep walking just because the map says there’s a bridge there.
I’m not saying that maps aren’t helpful. They’re incredibly helpful. But when you’re driving to dinner, I would recommend not keeping your eyes glued solely to your GPS. Checking out your windshield for oncoming traffic is probably a smart idea, too. (An analogy that I must credit to Gord Hotchkiss.)
This is the same healthy skepticism that marketers should bring to their use of data. Data isn’t always true, and even when it is, it’s almost never the whole truth.
Let’s Not Be Rash
I know, this may not be a popular point of view in an “Analytics” column. And I should probably talk to my doctor about all these itching sensations.
So let me reiterate: data is a wonderful thing. The intelligent use of analytics is an incredible skill for a marketer to wield — and an immensely powerful asset for an organization to have within its ranks.
But recognizing the limitations of data — and the cognitive biases that arise as much from using data as they do from not using it — will make you a better analyst, a better marketer and more fun at parties.
You can love your data. Just don’t fall under its spell.
Comments