The most successful ad campaign that I ever designed only ran two weeks before the owner of my company killed it. I was working for a software company at the time, but our parent company also had a consumer hardware division. I was asked to help with their advertising. The budget was extremely limited – every week we ran a 2×3 inch ad in the Saturday newspaper. That was it. Usually, those ads were crammed with data on different PC configurations that we were trying to push.
So one week, instead of advertising a bunch of specifications with a price, I ran an ad that said:
Why Pay Less?
I can’t remember the exact wording, but underneath that there was a short tagline that said something like “we have the best quality PCs in town – you get what you pay for.”
The next week, our sales doubled.
I ran the ad again the following weekend, and sales went up again.
Then the owner of the company killed the ad. Why? He didn’t like it. His argument was that people wouldn’t respond to it. When I showed him the sales data, he didn’t care. He didn’t like the ad, and that was that.
It was as though we ran a focus group of one, and he was the only member. I’ve worked for a few organisations where the CEO often made “Focus Group of One” type decisions. I’ve always found it pretty frustrating, because I’m more of an “Experiment, then Assess with Data” kind of guy.
Here’s a good talk by Erik Brynjolfsson of MIT, explaining how to use data to support innovation:
He makes a couple of important points. One is that the gap between companies that are performing well and those that are struggling has increased dramatically over the past 10 years. In his studies, Brynjolfsson has found that much of this difference can be explained by differences in the effective use of IT in enhancing innovation. Specifically, two of the areas where the better companies are particularly effective are in experimenting, and in diffusing the ideas the work well.
He uses the casino company Harrah’s Entertainment as an example of a firm that has done this particularly well (there’s a nice slide show on Harrah’s data mining here).
Harrah’s used customer loyalty cards to compile a huge amount of data on their customers. Once they had initial data, they then ran multiple experiments on innovative new methods for promoting their casinos. The data that they gathered enabled them to determine which promotions were most effective and which were less good.
One tangible outcome of this came in identifying where their money was coming from – over 80% of their revenue was coming from about a quarter of their customers. Those must be the high rollers, right? That’s how all of the casinos were acting, since it was the high rollers that were getting all of the free rooms and other stuff.
The only problem with that strategy is that those weren’t the people that were actually driving revenue. Instead, the money was coming from middle-class, mostly retired slot-machine players.
Once Harrah’s figured this out, they were able to develop a number of promotions targeting this group. Then they ran experiments to see which worked best.
It would be obviously much better to use tools like this to support innovation that generates a better world in some way, rather than producing people that gamble more. You need to make sure that you are innovating with purpose. However, the general principle of using data and experiments to improve innovation is still useful.
Using data and experiments is one good way to get around the “Focus Group of One” problem. It’s not the only way – there are many other approaches that can also do this, such as customer-led innovation – but one way or another, if you’re going to innovate, you have to figure out a way to gain a deep understanding of what your customers need.
And often, when you learn this, it’s not what you expected when you just thinking about it on your own. That’s the danger of “Focus Group of One” type thinking – it can lead you very far astray.