Being data-driven means being context-driven


It’s the current big trend. Every business claims it want to be “data-driven”, in other words  rely on data for anything. Is this disruption that big and why do they insist so much on this ?

Businesses have always been data-driven

As surprising as it may be, businesses have always been data driven. Or at least numbers-driven. Mainly on performance and profitability indicators. Believing or making believe that instinct and intuition have been prevailing in decision making for one century is surely an easy way to sell new approaches. The funny thing is that the main reproach made to these businesses is that they were looking at numbers only and used to make biased decisions. So, no. Data is not something new and what is shown as a revolution used to be seen like a bias yesterday.

When raw indicators embrace contextual information

That’s when Big Data shows up. But notice that people say more and more data than big data because, when it comes to data, we’ve always done as big as possible. Just remember that in the mid-80s a PC with 123 ko of RAM and a 5″1/4 floppy disk was a high performing machine and that today’s smartphone are incredibly more powerful than the computer used in the Apollo ship that brought humans on the moon. In terms of data and processing we’ve alway’s done as big as possible, but today’s big has nothing to do with yesterday’s.

Storage and processing capabilities help us to do radically new things in a radically new way. It comes with two main improvements. The first is we can use a wider scope of data in terms of quantity and variety. The second is the ability to process these data to find correlations and patterns at scale.

In the end it helps us to go beyond raw indicators and integrate their context and what impacts them.

Some examples.

Regarding sales management we used to focus on the signed deals and the pipe. Sometimes in link with marketing effort but these functions talk so little together than it was complicated to manage them jointly.Today we can take into account a large amount of data like weather, exchange rates, HR factors linked sales people and as many other sources as you want if they’re relevant to your market or industry. It ultimately helps to make decisions regarding to production, pricing, promotions etc. by relying on week signals that can be found light-years away from your business.

Same in HR to anticipate and avoid attrition. The risk of having key employees that leave can depends on management and HR data that are known and measures but also of contextual data that are outside of the HR scope. For instance it’s been demonstrated that many employee leave when their close ties leave and that close ties were not always they people they’re working with but those they interacted a lot with on the enterprise social network.

In short the power of data does not lie in raw indicators that are often simple and well-known but in the understanding of context, of exogenous factors that have nothing to do with the business but impact the indicator. A kind of butterfly effect.

The value of data is not in improving existing indicators but to provide new ones that help to make sense of the old ones.

It may look obvious but then we need to draw all possible consequences from that. Businesses will hae to collect data out of their original scope, sometimes out of the company, of the industry and of the business world. So they must be wise car processing any existing data in the world is not possible : they’ll have to decide what to take into account and improve the system over time.

Data is not the end of judgement

That said, we should not thing that moving to a data-driven approach will mean the end of personal judgement. As this article for McKinsey says, data and humans are complementary. As powerful as it may be, data only helps to understand a part of the context but does not provide an human story and narrative. ([What] I don’t think the data did do, was [have] an understanding of human story and human narrative. The data isn’t going to produce that for you”.). The solution is neither  human expertise or data but how you articulate both.

To some extent, I’d say that in a data-driven world, jobs like marketing or even HR will be less and less sciences or rational practices and more and more arts.

At this stage I still have questions when I see a business claiming it want to become data-driven.

• Is it a way to look trendy by using a popular buzzword ?

• is it a way to tell markets and investors on the business-orientation of digital projects since the experience-driven one are seen as qualitative and lacking real ROI and metrics ?

•a desire to remove biases from business decisions (even if it means overlooking common sense ?) or, at least, to fake it ?

There are lots of talks about deciding, much less about understanding. Poor articulation or propensity to overlook a major part of the contribution of data to business management ?