Careful what you look for: why generating value from insights is harder than it looks

I’ve been working in the data space for almost a decade now, and since I started, insights have been  touted as one of the most important contributions an analytics project can make. Now I wholeheartedly  agree with this, but it’s by no means a simple feat. Grand analytics projects often crash and burn due to  difficulty in value capture, and the lack of will to execute on the relationships uncovered. 

To exemplify what I mean, let's look at an example: in Recognition we work towards the retail segment,  and our experience is that retail is full of insights presented as truisms which are incredibly hard to  confirm or deny. One of the most common ones we’ve heard is “We put the milk in the back of the  store, and since everyone buys milk, they will be exposed to as many goods as possible, thus increasing  the average basket size”. It was probably true at some point, and in Norway practically all grocery stores  have this in their layout. Norwegian milk consumption has halved in the last 30 years, and only a fraction  of it can be explained through increases in dairy-milk substitutes like almond milk. This points to  changing trends like increasing veganism, or simply that milk isn’t as popular anymore. That seems like crucial insight worth considering when designing a store. If what I previously assumed was one of my anchor products simply isn’t, surely, I could replace it with something else? 

This situation highlights an important facet of insights: it’s dynamic and ever-changing, you will rarely  uncover insights that stay static for a long time, and it's not often you have the luxury of waiting 30  years for the changes to take effect. This is especially true if you act on the insights you’ve found, as you  are fundamentally trying to affect human behavior in some way, meaning one change leads to another. 

In some cases, the insights you gather might fundamentally change how you do business, which can be  both hard to admit, and to act upon. One of the issues with how insights are generated is that it is often  done as a one-off job, meaning there is no continuous data capture and analysis. We believe there are  multitudes of examples like this in retail, if you want to see how we help companies generate valuable  insights from data, check us out at www.recognition.no 

_______________________________________________________________________________________________

Brede Bjørhovd is the CPO of Recognition. Recognition delivers customer behaviour and traffic insights  to retail stores and malls at scale. If you want to learn more, you can meet Brede and Recognition at the  expo by visiting our stand ST193, coming to our presentation at the expo 1st March, 15:30 @ Theatre 7,  or check out our website: www.recognition.no.