One of the most popular topics in modern marketing is data, but for most marketers, the same fatal error prevents the potential data holds.
In his book, Thinking Fast and Slow, Nobel Prize-winning economist Daniel Kahneman, one of the founders of behavioral economics, discusses a bias he refers to as ‘The Law of Small Numbers’. Without getting into his entire thesis (although it is fascinating and worth a read) the idea behind this law of small numbers is that we have a particularly difficult time processing very large sets of data. It is for that reason that we tend to come to conclusions using smaller, less accurate data sets. If you want proof of why that’s a problem, just look at the 2016 election. Small, statistically relevant sample sizes, said one thing would happen, and something else happened entirely. Of course, sets being statistically relevant matters when it comes to analytics, but today there is a lot more data available, and it is much more easily accessed; we just need to jump over the hurdle of wrapping our heads around large data sets.
The Law of Small Numbers
Brands, advertisers, and marketers love to make generalizations. If ten people say one thing, that means everyone will say the same thing. If a room of strategists agree on a plan, it is sure to work, right? Obviously, samples and expert consensuses are great, but the law of small numbers suggests that small data sets and sample sizes lead to sweeping generalizations. Sometimes those generalizations hold true in practice, and establish themselves as proofs (though after the fact) but while we tend to only hear about campaigns that succeed, there are far more that go unnoticed and forgotten, in many cases because the pre-conceived generalization did not hold up.
Despite the availability of mountains of data across multiple media, the world of marketing still tends to do things the old-fashioned way. Sure, we hear a lot of buzzwords and plenty of campaigns reference the use of data, but when it comes to the uses of troves of data in a predictable model before a campaign is developed, the case is often that the use of data is overreported but underutilized.
Preventing Big Mistakes
When we start with our owned data and insights, then expand from those insights to apply and contrast what we learn to a larger audience, we can begin to avoid the pitfalls of generalizing. Instead of focusing on what has worked, we need to recognize that tastes and preferences change, and that all of that information is available in large data sets. It can be done, we just have to move past our distaste for large, slightly more complex amounts of data.
Latest posts by Corey Padveen (see all)
- A Common Fatal Error Most Marketers Make - November 29, 2017
- Rising Utilitarianism in Decision Making: Cause Marketing - August 24, 2017
- Rising Utilitarianism in Decision Making: The Sharing Economy - July 20, 2017