Big data needs small data

12 Apr 2016 by Scott Middleton

Statistics, numbers, big data and artificial intelligence are taken far too often, consciously or unconsciously, as the best and only answer. Now that we have such powerful information at our finger tips we need to remind ourselves to get back to thinking small – focusing on the tiny clues our customers, suppliers and the market leave that reveal immense opportunity.

Lego, now a beacon of innovation and growth, was once, according to their CEO at the time, “a burning platform, losing money with negative cash flow, and a real risk of debit default which could lead to a break-up of the company.”

Martin Lindstrom, author of Small Data and a consultant engaged by Lego at the time, says Lego had lost its way due to an over reliance on big studies forecasting trends towards digital and away from the fantasy world of Lego blocks. The studies drew on things like well-known socio-economic trends and forecasts on the impact of technology like video games.

The turnaround came when Lindstrom and others met with 1 – just one – 11-year-old German boy. Lindstrom says the boy helped them realise that Lego “was all about the summons, the provocations, the mastery, the craftsmanship and, not least, the hard-won experience.” Lego’s focus now sees them as a leader of innovation and one of the world’s largest toymakers, if not the largest.

Lego isn’t alone. We’re all easily influenced by numbers, especially those presented as statistics based on large (or even small) data sets.

One key takeaway from Nobel Prize winning psychologist Daniel Kahneman’s decades of work is the ease with which people unknowingly arrive at decisions due to the ways in which data and information are presented.

Kahneman’s book Thinking, Fast and Slow provides numerous examples of studies that demonstrate how statistics aren’t what they immediately appear to be.

One interesting example is a study on the incidence of kidney cancer in the 3,141 counties of the United States. Kidney cancer was found to have the lowest number of incidences in a few states, sparsely populated counties. Misinterpreted, this may result in leaps to conclusions about how healthy these states are due to their rural nature and, who knows, investments may be made into trying to understand how these states are so great at preventing kidney cancer.

The brains nature makes it desperate to find causality however Kahneman tells us that sometimes there is no cause and the laws of probability have led to this statistical outcome.

So continue to expand your data collection and analysis. Just be sure to explore a bit below the surface and question the results.

Back to Blog