As long as nothing gets done

?I have been working with data since a while, mostly in Business Intelligence or more widely in Data Driven Processes (creation, management, optimization...).

There are many things in common in the two subjects, one, in my opinion is KEY: the end game is to perform an action.

Today I was participating in an interesting online event in which the speaker showed a slide that looked like this:

Data → Information → Insights

If you are familiar with this subject matter you probably know that that slide a few years ago would have been:

Data → Information → Intelligence

What’s the difference? Why did things change? Is it an improvement?

The beginning of the cycle is the same and can be summarized in this way :

Data describes something that happened (facts) and its context (dimensions)

Information is the next step and is generated by the understanding of what happened

While there are indeed many challenges related to these two first steps (data: availability, quality, ingestion, transformation… - Information: business rules, correlation, causality … ), it’s the third step that usually is extremely hard to achieve… or at least it used to be.

Intelligence in this context means doing something with the information we have, that’s not what Insights means.

What does “insight” mean?

Insight /??ns??t/ noun

"the capacity to gain an accurate and deep understanding of someone or something."

So, after Information which implies understanding/learning now we have some deeper understanding… but we don’t necessarily do anything with it.

Imagine the Central Intelligence Agency becoming the Central Insight Agency, not sure it would serve the same purpose.

So why this current shift?

Convincing ourselves we can do things is something we learned to do quite well, actually doing them proved much harder.

How do you measure the ROI of your knowledge? I’d say you test it when you apply it in some process, choosing path A instead of B because of it.

If we stop at “Insight” then we may believe we can easily make the right choice, but there is no implication or requirement to actually make it, hence you will more easily achieve your goal.

To me, the biggest difference with the two approaches is that Insights might not be relevant while intelligence is always, by it’s nature, specific to a need.

If my problem is something like: “Should I buy product X or product Y”, only facts, attributes, rules etc that are relevant for that specific decision are relevant. This is how we build intelligence: we start from the problem at hand and walk our way backwards to information and to data.

On the other hand if you collect all the data you are able to collect, you put it in some sort of repository and see what you can discover from it, you potentially get a lot of insights, some might be useful, some might not, but nobody will ever know… and this is how you achieve a project.

Don’t get me wrong, I am not saying that insights are “bad”, what I am sayind is that they should not be the end game.