Why we should be data-conscious, not data driven

Michael Hoole
Feb 25 · 5 min read

How having a clear understanding of your organisation’s values stops you from drowning in data

What is being data-driven?

Data-driven is a phrase we hear a lot these days (and I’m definitely guilty of having used it myself) and it tends to be followed by lots of grave head nodding like this is self-evidently a good thing.

“We’re data driven. We cut through the noise of emotions, feelings, and hunches and we now make intelligent, informed and data-driven decisions”

Is this an accurate reflection of what really happens though? Is it even something that we should be aspiring to? I’m not so sure that it is.

How do we actually use data?

Let’s run through an exercise.

You are the Editor-in-chief of a prosperous news media company. Your company prides itself on being a ‘data-driven organisation’. You are presented with a list of five articles and must decide which one gets the top promotional slot on your website. This slot gives the article 500x more views than the average article so it’s an important decision.

A list of five articles numbered 1 to 5 in descending order of page views
A list of five articles numbered 1 to 5 in descending order of page views

So which one of the above articles do you choose? My guess is that most people would probably go with either Article 1 or 5. You’re either choosing Article 1 because it seems to be the most popular and therefore should be the front-runner for the top promotional spot or Article 5 because it’s the least and needs a boost. In this sense we can say that we are making a ‘data-driven’ decision. But then what if we add a new data field…

The same articles listed, now with ad revenue per page view included. Articles 2 and 5 have the highest ad revenue per page view
The same articles listed, now with ad revenue per page view included. Articles 2 and 5 have the highest ad revenue per page view

With the ad revenue data added, suddenly the decision as to who gets the top spot becomes an easier one to make from a profitability point of view. Articles 2 and 5 suddenly seem the most attractive options. But then what happens if we add yet another column of data?

The same articles, now with audience response percentages in categories ok, good and great. Article 4 has the most positive audience response.
The same articles, now with audience response percentages in categories ok, good and great. Article 4 has the most positive audience response.

Each article has feedback buttons located at the bottom to understand what readers thought of the piece. Suddenly making a data-driven decision increasingly resembles a value-judgement. The more profitable articles have the least enthused audience response. Article 4, which had neither much in the way of existing page views or ad revenue generated, is the one that appears to be connecting best with the audience. So now you have a trade off between engagement and profitability. Knowing only that your organisation is ‘data driven’ is of little help in guiding which article gets that top spot.

Why values matter

So what if you are then told that your news media organisation has built its reputation on producing and promoting highly engaging content to a wide audience? Which article do you choose now? It has to be Article 4 right? Knowing more about your organisation’s values, the ‘Audience Response’ data now becomes the field that you would most likely base your decision on. So what about the other data fields? Do we just dismiss them? Of course we don’t — they may not be driving our decision-making but they are invaluable in our wider data-consciousness. They prompt questions like:

  • Why is our ad revenue so low for our most engaging article?
  • Why has it received so few page views to date? Is there something wrong with who we’re promoting it to or where it’s located onsite?

These are real questions we can ask that will help inform future strategy even if they are not driving our immediate decision. This is what I would call data-consciousness.

What’s wrong with being data-driven?

The problem with the phrase ‘data-driven’ is that it treats data as an apolitical entity that informs value judgements rather than being subjected to them — when we know this isn’t the case in our lived experience. Who collects the data, what data gets collected and which data is valued most highly are big, complex, morally ambiguous questions that we shouldn’t (and can’t) shy away from. Pretending that data guides our values rather than the other way round is faux-scientific and, worse than that, leads to conflict and time wasted as different stakeholders use data to justify their own agenda. We can all find examples in which data has persuaded us to change our actions — it’s much harder to think of when it has changed our values. That is why our political beliefs tend to be fairly consistent even when presented with data that promotes an opposing viewpoint.

Another issue with being purely ‘data-driven’ is that we have a natural tendency to lean toward data that is a) easy to collect and b) easy to influence. This can lead to shallow insights or focus minds toward trying to influence superfluous metrics. Customer research tells us that FT readers value our content for our perceived integrity and the uniqueness of our content, yet these data points are far more difficult to capture despite possibly being the FT’s most valuable assets.

Becoming values-driven but data conscious

Values give us a framework for what data we collect and how we interpret it. Values help us align across various departments and to make decisions quickly and effectively. Perhaps most importantly values stop us from drowning in data. Aiming to be a values-driven but data conscious organisation seems a far more honest and meaningful aspiration than that tired old trope of data-driven. The first step we can make toward this is to stop using that phrase in our conversations.

FT Product & Technology

A blog by the Financial Times Product & Technology…