Regression to the mean(ingless)

Frederic Kalinke's eConsultancy article "Marketers have more data than ever, so why aren’t they better at experimentation?" should be a gateway drug to actual science for marketers. (It's also great because it starts with a Paul Feyerabend quote).

I say "actual" because our there is a significant opportunity for marketing (and business in general) to learn from real science, so we are less likely to draw conclusions from data that do not (or cannot) lead to good decisions. Kalinke's article touches on a number of pain-points that we all face in trying to come up with a specific solution to a specific opportunity that is highly contextual, possibly edge case in nature, or demanding innovation, but by relying on insights or assumptions based on de-contextualized, aggregated and normalized information.

Carl Jung's book "Modern Man in Search of a Soul" illustrates this by describing a beach full of pebbles, the average weight of which is 3 grams. Just because the average weight is 3 grams, doesn't mean you can expect to reach down and pick up a 3-gram pebble. In fact, a 3-gram pebble might not exist. Because what is being measured is something that exists in between all of the objects, but is not (or may not) be a property of any of the objects.

When you are creating content and customer experiences, you need something highly tailored to your audience, especially for a perfect representative of your audience.

As Seth Godin sayeth, you need to target the smallest possible audience with your offering, not the largest. Something that appears to be counter-intuitive for those accustomed to 'broad targeting' (a quasi-oxymoron), or who are operating on assumptions based on multi-causal multi-year post-synergy success stories, in order to make decisions about how to be causal, in an early-phase, pre-synergetic context.

To go to Godin's illustrations, Nike's targeted athletes, AirBnB's hand-tailored and11-star thought exercise driven model, and Apple's initial niche community focus, and many other companies didn't start with mass-targeted products defined by averages, mass or diversified segmentation. They started with niche products, highly tailored to the specific needs or a narrow or ideally perfect audience. Once their products gained traction there, then they were able to evolve and scale to broader targeted and more diversified success, building on a niche foundation.

The value was in the niche. At the edge. For the ideal. Outside of the echo-chamber of the average.

Whenever you hear someone say "we should move to video as 35% more time is spent on video that last year"—accross all media use-cases under the sun—you are hearing a decision or direction driven by the wrong data, based on the wrong questions, applied in the wrong way.

Aggregate numbers can create profiles of broad groups (a sample, a geography, a slice in time). They can help us ask better questions. They can input to our hypothesis to be tested. A hypothesis based on an average can be tested against a hypothesis based on an edge case.

But this assumes a testing regimen. All too often directions are given and options are prematurely narrowed based on someone quoting an industry statistic from a trusted consultancy brand. This is especially true if the opinion is expressed by a HiPPO (Highest Paid Person's Opinion) and teams fall all over themselves building the leader's echo-chamber from the bricks and mortar of shareholder dollars.

When looking for what is meaningful, you have to get down into the soul of a person. Into the dreams (and nightmares) definitely shared by a well-defined target group. Something not one, but all of them will nod their head about at the slightest inference. Something bound to the code of the tribe. The absolutely common enemy. The absolutely shared dream. It will be a small number of people. It might just lead you to something never done. Or at least, to something that someone will actually appreciate what you did.

This act, will probably veer from the mean, and the meaningless.