Quality, not quantity. Less is more. Remove is better than add.
There are many expressions that, accompanying us in our daily life, praise the small, the selected, the accurate. Equally numerous are however, in today’s business, the devotees to big data, to the analysis of which they seem to entrust the answer to every question arising in modern era. According to Martin Lindstrom, brand consultant and marketing guru who will lead the BBS workshop Data Driven Strategic Marketing on November 29th, big data provide an infinite amount of impersonal information used to predict future business and brand orientations, but only the individual and unique small data coming from individual human beings, can lead to a true understanding of reality. Actually, computers and software are only as good as their best programmers and to see the big picture, there is still a need for human eyes.
Big data are typically used to understand the correlations between large data systems, find dependencies and predict future behaviors. At the same time, however, according to Martin Lindstrom, at least 60% of the biggest inventions of all time – like Post-it or Snapchat – are based on small data. In his book Small Data – Tiny Clues that Uncover Huge Trends, Lindstrom explains the surprising origin of the intuitions that are at the base of some of the greatest successes of global brands.
Lindstrom’s interest in small data and branding at large started and developed through LEGOs and the direct contact with the company that, attracted by the Legoland that he had built in the garden as a kid, enrolled him as a builder and innovator at the age of 14. In 2004, when the company was on the brink of bankruptcy, all the studies based on big data suggested that their kits should be simplified to make them more suitable for digital natives who, according to the consultants, had neither enough patience nor imagination to play with bricks. But it was the visit of the LEGO marketers to an eleven-year-old German boy to overturn the fate of the company. The boy, proudly exhibiting his worn out sneakers, explained that they were a proof of his dedication and skill with the skateboard. At that point the LEGO team understood that kids wanted to show to their peers that they master a certain skill and that the LEGO kits, instead of being simplified, had to become much more complex.
Similar is the story about the robot vacuum cleaner Roomba and its abrupt halt in sales. With the help of Lindstrom’s suggestions, the company found out that the decline in sales was due to the modification of the navigation sound, which from funny exclamations turned into electronic beeps, and to the revisiting of the design. The company understood that people were treating Roomba as a pet and a surrogate to human company and that the current changes made it less ‘alive’ and more similar to a home appliance. These data, impossible to extrapolate from the analysis of big data, were fundamental to rebuild the relationship of trust between the company and its customers.
According to Lindstrom, big data need a hypothesis to be tested and verified in order to become fully useful: “You can not simply look through millions of data and come to conclusions. Walmart, for example, has the largest data center in the world, is twice those of CIA and FBI put together. The price of its shares nevertheless fell by 11% in 2015 “. If we all have access to the same data and analyze them with the same metrics and the same softwares, we come up with identical solutions. Which often turn out to be incorrect.
According to Lindstrom, the analysis of small data is necessary to fill the gap left by the two main limitations of big data: the inability to provide new ideas and the emphasis on analysis at the expense of emotional factors.
By their nature, big data have, and will still have in the future, the ability to connect large volumes of data of different types and sources, thus giving the company a look at the outside to understand trends, identify potential customers and formulate new value propositions for them. The big data, in short, explain why a phenomenon is taking place, while the small data try to intercept what is happening, focusing on the details. In a sense, big data explain the motivations of phenomena already underway, tending to consolidate the results and recommendations around what we already know, replicating patterns and proposals without leaving the necessary space for new and different solutions.
Strong brands can count on two advantages: the respect and the emotions they evoke. Big data, however, by giving priority to analysis to the detriment of emotional cues, can not describe the emotional qualities to which individuals attribute more value. If it’s true that brands like HP and Duracell are based on respect and big data can definitely help them make concrete decisions about how to increase it, brands like Disney and Cheerios are based primarily on the feelings they are inspiring and this kind of consumer relationship can not be stimulated by the analysis of huge quantities of anonymous data.
To extrapolate information from big data also requires a technological effort and skills that especially small businesses are not able to deal with. Several companies are therefore understanding that in some cases, similar results can be obtained using much less structured and complex data mining strategies. Contrary to big data, in fact, small data are sets of very specific and circumscribed attributes that often provide ready-to-use and timely information, organized in an accessible way for companies to understand the end customer, his needs and actions to take to reach him.
Although big data can connect millions of information to identify the correlations, their effectiveness is compromised when they have to deal with the complexity of human beings. One example is the driverless car developed by Google that did not manage to deal with a four-arm crossing because its sensors were calibrated to wait for the other drivers to stop completely, but it was not programmed to interpret and interact with a world where the rules are not respect. Without a doubt, big data will help us save time and money and automate life, but in parallel humans will evolve to circumvent the new problems created by technology and, in many cases, to deceive it.
A recent qualitative research conducted in Switzerland revealed that almost all of us have up to ten different interdependent social identities: we are parents, friends, the funny ones on Facebook, busy professionals on LinkedIn, employees, amateur runners and much more. The role of anyone trying to make sense of small data is to understand not just one of these personalities, but all of them.
But how can small data overcome the sampling bias? And how can a company hope to find a solution or a definitive answer in such a small sample? “My answer is that a single drop of blood contains enough information to detect almost one thousand different virus strains,” says Lindstrom. “More difficult to admit, for many people and companies, is that sometimes there is no need to carry out research on millions of consumers, but ten people are enough to transform a brand or a business.” Many times, in fact, an entire population can be explained by a small signal that acquires meaning only if correctly contextualized.
“I had never considered my work to consist of a replicable methodology, but in recent years, several companies have asked me to systematize my subtext research to develop a training program. In Nestlé, for example, my method has become an integral part of the analysis of new products, ideas, innovations and brands. Today, thousands of Nestlé employees spend 48 hours a year visiting consumers in their homes,” adds Linstrom, who gave a method and a systematic approach to the observation of consumer behavior and emotions. “After all, at the age of fourteen, when I joined LEGO, I was a consumer, just a kid in love with bricks. By observing my behavior and that of my friends, I was able to offer LEGO executives information about their product and their company that they could never get from quantitative surveys: and in the same way, in stark contrast to what big data said to them, the observations of an eleven-year-old German had helped LEGO to save itself from bankruptcy.”
Although the small data represent a sort of inversion of direction compared to the increasingly widespread obsession with big data, it would be wrong to believe that they are losing relevance. At the same time, however, if companies want to understand consumers, big data offers a precious but incomplete solution. Despite all the useful information that the web provides, the data we generate and leave there remain a curated and idealized version of who we really are. The future integration between the two perspectives is not only the most probable solution, but also the optimal one, where the gaps of one approach can be filled by the other.
“Small Data are not concerned with testing concepts but they create the basis for innovative brand thinking.” – Martin Lindstrom
For more information on Martin Lindstrom’s workshop Data Driven Strategic Marketing, which will be held on November 29th at Bologna Business School, please visit the dedicated page.