The Math Trap

Too Much Data Can Strangle Creativity
September 12, 2018 Omnibus

Photo by Tyler Hendy on Unsplash


Some of the brands surrounding us are more than 50 years old. Some even more than 100. Think of the good old international brands like Coca-Cola (1886), Nike (1964), H&M (1947), Milka (1901), Nutella (1963), Wrigley, Porsche (1931) or take more regional angle with Radenska (the three hearts logo created in 1936), Domačica (1954), Mikado (1964), Cockta (1953), Barcaffe (1967), Elan Skis (1945). All of them – still well & vividly around – came into being in the direct interaction with the users and customers.

Back then there were no intermediaries. No big data. No market research. The development of some of the above brands leaned more on the refined taste of the producers, the others more on the feedback of end-users.

Says Mrs. Marija Tul, nicknamed “Coffee Mom”, how the brand Barcaffe – still by far the most popular coffee brand in Slovenia – was born back in the 1960s.

“At first I tested the flavor alone. Then we checked it out with our fellow co-workers. Those who sensed the really tiny content differences in constituents, like sugar, acid, salt were added to the tasting sessions.  We always offered more refined taste according to our team preference, so the customers had to adapt to it.”

The coffee smells too much of a socialism? Perhaps try another one.

How many focus groups did Mr. John Pemberton address to set his beverage into pharmacies – the product that allegedly cured against morphine addiction as well (the man was addicted with morphine himself)? The drink was named Coca-Cola, of course.


Later market research entered. And customer segmentations. And demographics. And more segmentations. And more data about past sales. With big data, we fully entered the time of the metrics & the time of the analytics.

Of course, the companies are on the different levels of the data use:

1. Some rely on sales data analysis alone.
2. Others add all sorts of customer segmentations (taken from loyalty cards, market research etc).
3. The more sophisticated have entered the field of Big Data – adding the patterns of behavior, mining the data, employing the neuromarketing research etc.

In the end, by definition, when all the big data about customers is collected and processed – this should mean more engaging products/ services / formats / brands.

But is it really so?

Take a look in the retail arena. Yes, there is a constant flow of new products in the supermarkets but are we really getting more and more excited about them? No. On the contrary.

In reality, we are encountering blander and blander products gaining lots of promotional space while the most exciting new entries are knocking on the doors from the side, usually creating its own category at the same time. Think about energy drink Monster, Innocent smoothies, Amazon, Danish fun-inducing retailer The Flying Tiger, and other brands – all actually stemming from the independent scene, some even from craft scene.

We can move further to other arenas. Pretty much the same. Just recall the designs of today’s cars. Where are the bold if overblown statements of the previous times? Yes, sometimes visually pleasing but not actually engaging. Stirring no emotions at all.

So how come that all the Big Data doesn’t provide the new exciting products that would really sweep us from our sofas and find the direct way to our hearts like the ones listed in the first paragraph?

The answer might have a lot to do with the math trap.


The math trap is something that happens when you misidentify the tool with the end product. The data analysis gets you so overwhelmed that you overlook the most obvious: The past data is not the best indicator of future expectations.

In the Josef Adalians revealing behind the scenes article Inside the Binge Factory  Ted Sarandos of Netflix says:

»You have to be very cautious not to get caught in the math because you’ll end up making the same thing over and over again.«

And let us pause here for a moment. Isn’t Netflix the company that in the past 7 years established the new business model built exactly around its elaborate Big Data algorithms?

Right. That comes from the Netflix creator of such successes as Stranger Things, House of Cards. The one Netflix being an instrumental player at the core of the Golden Age of the TV Shows. And these producers sure know something about Big Data. When creating new products, they leverage the immense databank, consisting of more than 2000 different patterns of customer viewing behavior. Wouldn’t be an exaggeration to call Netflix champions of the Big Data.

So, why – all of the sudden – Netflix denies the very basic principle it has brought them so many successes?

Because they know it better. They escaped the fault of mixing the tool = Big Data analysis with the final goal of the product development.

Adds Ted Sarandos of the Netflix: “It’s 70 percent gut and 30 percent data … “Most of it is informed hunches and intuition. Data either reinforces your worst notion or it just supports what you want to do, either way.”

What? The guy who controls $6 billion of yearly budget rooting for the intuition and that gut feeling?

Yes. Because the making of the successful product is not (only) mathematics. It’s not even (only) science. At the right moment, something else comes into play. Might be called imagination or just the core of the creative process (intuition or the gut feeling).


Also in the offline world, take retail, for example, we need more Netflix approach. Yes, employ big data, use it smartly but then – draw the line. Stop! Stop analyzing, start creating.

Pure past data might be good enough for those who just want to copy – capitalize on something already done in other markets. But when you want to really engage customers, you have to explore, to venture into unknown.

Creating something new – a product, a service, a store format, a TV show – is like a drive down the roads less known. Checking the rearview mirror (past data) is important to evaluate the situation around you. But the rearview mirror (past data) is pretty much useless for taking the right decisions about the direction at the approaching crossroads. This is the moment the good old signpost steps in – intuition.


Leave a reply

Your email address will not be published.