It is time to put judgement back into our analysis

by Nigel Hollis

When I first started work in the U.S., I vividly remember being taken to task by a senior research director at a client company who accused me of using “judgement” in my analysis. This episode highlighted a clash between my belief that data needed to be interpreted to be useful and a widespread belief that the research role was simply to report the data.

Of course, since then the industry has acknowledged that “insight” is the name of our game, and that we ought to be interpreting what data means and recommending what action to take. Except that I am not sure anything has really changed that much. Otherwise the industry would not be clamouring for real-time, sense and respond data on everything, regardless of whether it makes sense or not. There are times when you need to act immediately and there are times when you would be better served by taking more time to really understand what is going on.

I suspect the root cause of our obsessive focus on immediate optimisation is just the same as it was when that research director accused me of using judgement. That cause is a belief that the numbers should tell us what to do. The research director wanted a statistical proof that a new ad campaign had indeed lifted brand awareness and trial – irrespective of the obvious eyeball relationship – and today we want the apparent surety that big data, AI and machine learning bring to the party.

There is no doubt that the analytic techniques available to us today offer amazing opportunities to make decisions on the fly without recourse to human judgement. Take the example of Nvidia’s self-driving car that taught itself to drive by watching a human do so. As Will Knight’s article suggests this is an impressive feat, if somewhat disconcerting because there is no way of knowing how it makes the decisions it does. But perhaps more important is the fact that without someone to tell the car where to go the technology will be useless.

I would argue the same is true when it comes to marketing. My analysis of the less than one in 10 brands that grew over a five-year time frame finds that they all did something disruptive, something that changed the way they served customers, went to market or communicated with their target audience. They created something that did not exist before. Understanding your customer, anticipating their needs, wants, and desires, is critical to achieving growth but it requires judgement to identify the best strategy to do so.

This is where I believe we need to take more time, to delve into how people use a category today, how they think (or do not think) about decisions, and figure out how to change it to favour a specific brand. While big data, AI, and machine learning can play a valuable role in that process, I believe human imagination and judgment will still be required to envisage what could be and not simply respond to what already exists.

Insight into what people know, do and say now is useful but the real competitive advantage lies in figuring out what we can encourage people to do tomorrow and that might take a little judgment.

 

Skip to content