The fallacy of being data-driven
Power to data
Data has always been essential to take informed and potentially correct decisions. Decisions that — for any company — are necessary to build medium and long-term strategies or to start short-term tactical operations.
Today, each of us generates a huge amount of data. Constantly, continuously, silently and sometimes not consciously.
Our purchasing behaviour, the sites we visit, the things we research online, the emails we read, the posts we comment on. If and how much we train, how we sleep, if we have an accelerated heartbeat. Where we travel, our travels, who we know and with whom we speak more frequently. And in some cases even how many washing machines a week we do or what the temperature is in our living room.
There is no doubt that companies today have not only the possibility of accessing a quantity of data never seen before but also of being able to do it in real time and automatically.
Data is the new currency.
You will probably have heard the phrases of Edward Deming “Without data you’re just another person with an opinion” and “In God we trust; all others bring data”. In courses at Pragmatic Marketing, a frequently used sentence is “Your opinion although interesting is irrelevant” meaning exactly the same thing: without data, without ‘supporting evidence’, any decision, intuition, opinion is completely useless.
The fallacy of being data-driven
This sudden availability of data combined with the increased processing capacity has meant that in recent years being data-driven has become something essential for any professional and for any company. All are somehow “data-driven” or — even more extreme — “data-obsessed”.
The adjective “data-driven” means “to advance in an activity guided only by the data, rather than by intuition or personal experience” [Wikipedia]
Fallacy seems to be an exquisitely pragmatic question (in the sense of studying the ways in which context contributes to meaning).
The emphasis on being driven by data or obsessed with data leads us to focus exclusively on data, often giving up to consider other factors that can make the difference (and often do).
If you’re a marketer or a product person, chances are that you’ve probably started doing more and more experiments that, given some numeric results, justified certain choices.
Maybe it’s when we started talking about big-data, that suddenly the data became the supreme decision-maker. Artificial intelligence, machine learning, predictive analysis. The analysis is driven by a huge quantity of data to say and tell us what to do. (Then came the small-data but this is another story)
Technology helps us a lot today. It helps us to obtain information, to memorize it, to analyze it. And this is a great thing. For those involved in marketing and product, in reality, the data has always been fundamental.
But perhaps the scarcity and the difficulty in analyzing it forced us — in the past— to carefully choose what to focus on and constantly question the meaning of a number or a result. The data provided information. Information that was used together with others to evaluate the best thing to do.
Today we have many tools available that make the collection process, analysis, data management very fast, real-time, automatic. And, frequently, decisions are made (in an equally automatic way) guided only and exhaustively by a single result. Which is often a short-term result and of which we know nothing about a longer timeframe.
And I think the problem is not related to the recognition of the importance and centrality of the data — in which I firmly believe — but in recognizing the limits as well.
Data can lead to invalid findings
In marketing more and more we focus only on the data to decide the target, the channels, the number and time of the posts, up to the definition of an entire strategy. And also in the product, all the recent methodologies require continuous experiments between version A and version B, where the version with the highest data wins. But data alone can lead to invalid findings.
It happens that I decide to publish my posts when I get the best click rate. Ignoring that — compared to 100 clicks more — I’m annoying 10,000 users who do not like to receive a notification or an email at 3 am.
Or I assume that my data is always valid and up-to-date. It matters little if I continue to send content that targets the deceased daughter of a mother.
Or I adopt a different style of communication, creating posts with click-bait titles and lots of buzzwords, to which I put some likes and comments from my network. If I do this the data tell me that I get more clicks. And in fact, I increased traffic to my page by 560%. I do not understand, however, why the target that most interests me has instead reduced enormously.
And that’s how I decided that the main colour of our new app will be red because that’s what 51.999154% of users prefer. It matters little if in our brand image red does not appear absolutely. And indeed, thinking about it, it is the colour brand of one of our competitors.
Data has always been and will always be fundamental. It is the exclusive importance that we often give it that should make us think.
Data and knowledge
At this point, I think my point of view is quite clear.
I do not believe that the data itself can or should guide any decision.
It can and must help to make decisions that must, however, be guided by a strategy of intuition, vision and knowledge.
It is precisely when we speak of knowledge that the data take and expresse their fundamental usefulness.
Knowledge is — in fact — the awareness and understanding of facts, truths or information obtained through experience or learning, or through introspection. Knowledge is the self-awareness of the possession of information connected to each other, which, taken singly, have a lower value and usefulness. [Wikipedia]
Knowledge is, therefore, something different (and of greater value and utility) from the simple data or information derived from it.
While data and information can exist independently of those who can use them and can be preserved on some kind of support, knowledge exists only because there is a mind capable of possessing it.
In 1983, the Russian satellite system gave the alarm signalling a missile launched from the Malmstrom base in Montana, travelling to the Soviet territory.
All the data were giving information: “a missile attack is underway”.
But Lt. Col. Stanislav Petrov had an intuition. How is it possible that America — which had thousands of missiles — began an attack by launching only 5 missiles? Such an action could not make any sense.
His knowledge and intuition prevailed over the data and the only information they provided. And so he decided not to start the Russian counter-offensive.
If Stanislav Petrov had been “data-driven” the world, at least as we know it, it would probably no longer exist.
Knowledge-driven
I would like that in 2019, the concept of “data-driven” becomes closer to being “data-oriented” than “data-obsessed”.
I know, it’s probably just a matter of saying so; but words are important.
My personal wish is to pay more attention to knowledge and become all a little more “knowledge-driven”.
First, because knowledge is something in which the human component is fundamental. And it is the one that still lacks artificial intelligence.
Secondly, because knowledge allows us to make counter-intuitive but winning decisions.
Recognize that the data is fundamental, but that alone is not enough.
Data is the new currency.
But it has always been the knowledge that makes humanity leap forward. The data, if anything, can help knowledge to jump farther.
