Chips, as healthy as raw cashews - NutriScore & the implications of single indicators
Universal color schema & data visualization
My son Leon, age 5, knows that in the grocery shop some cereals can have a "grade" - from letter A (green) to letter E (red) - and he knows that green is good and red is bad. This exists in some European countries and it is called NutriScore. So, when shopping, Leon suggests that we go for food the healthiest food possible, just by looking at a letter-color combination. And that's a very good example of how proper visualization can transform complex information into a clear, accessible, unambiguous insight. We are not going into the complex topic of saturated fats, processed fats, sugar content (ok, we do, for sugar it's easy), fibers, proteins etc - but instead, it is a single "KPI" telling a very young user what is the the better choice.
Universal KPI
Fast decision making at senior levels
Obviously, that's great as intention, in the context of cereals - more sugar, lower score - even a five-year old can figure it out. As a father, I should support it as it helps my son be more health-conscious. There is however, a deeper layer here - some food for thought or thought for food if you wish, especially if you are coming from the world of data: this is an example of a unifying KPI - one metric to supposedly try to tell you everything about the healthiness of a food. You can often hear the argument that we can't expect senior leads to look at all different nitty-gritty details, all of the 20 reports for the teams they're responsible for - they should be able to identify, within seconds, who is to take action, not necessarily the kind of action.
Hidden dimensions
Personally, I have always been a bit of an opponent of universal KPIs as, I believe, they are likely to obscure deeper levels of information and leave decision making process at a shallow(er) level. Further, there could be a hidden dimension in this data, a perspective, a caveat that remains unknown for the end user. It could be that the methodology of attributing weights to the different components doesn't include everything that is important for us or the prioritization is not the same as what we have in mind. It could be that these results have to be seen in a certain context to make sense. Finally, there could be another dimension missing to make the right choice. Going through the Wikipedia article for NutriScore we can see that indeed, these values are comparable, only within the same group of foods - that is why, it is possible to have deep frozen fish sticks with NutriScore of B and raw nuts with score of C. Our young data user can make the decision that if he wants to be healthy, he can munch sugar-free candy as much as he wants - as the score is "A". We can even see the paradoxical situation, where raw nuts and chips (crisps) have the same Nutri-Score. If Leon was to use other available information, such as price and volume, it's likely that he would have ended up with a bag of nachos...
So, this is my question for you: do you believe single KPIs can make life easier, by allowing for a faster decision making process, conflating multiple variables into one "Measure to rule them all"? Do you feel that end-users of reports should be able to see a single color or a letter to evaluate a complex process? Or perhaps you feel that the color would trigger them to take action and dive deeper?
Data Analyst
2yI'd say they can occasionally be fine, especially when the original intent is not corrupted by lobbying.
Board Member at Curacao Data Driven
2yThanks for sharing Vitali. You wrote: "do you believe single KPIs can make life easier....." as always the answer is: it depends!