Not rules, but tools

Imagine how powerful the reams of data that digital marketing generates could be if made available to all in one location. If only such a place existed

A vice-chairman in the UK went to see a talk by Esther Duflo, the French-American economist recently awarded the John Bates Clark Medal - an economics award often seen as a precursor to a Nobel Prize.

He came back dumbfounded by the talk, which he described as "Drayton Bird meets Mother Teresa". He hasn't stopped talking about it. And now he's got us very excited too (as well as giving us the perfect subject for this essay).

This is what got Rory (Sutherland) even more animated than usual. Duflo centres her attention on one question: "Why do the world's poor stay poor?" Helping the poor, she says, is "unambiguously the most important and interesting question. You can think about how to help very rich people make a little bit more money by improving the efficiency of tax collection. Or you can think about how to improve income and quality of life in Mali, where people are making $240 a year."

Nobody can argue with that. But it is Duflo's methodology that Rory said so surprised him, and we can see why. It's essentially a direct marketer's approach applied to the eradication of poverty.

Central to her approach is randomised experimentation, where researchers hone in on the impact of a particular variable.

In one test, she created control groups to see whether offering a free kilo of lentils to mothers for inoculating their children was a cost-effective incentive (it was). In another, she tested the value of free school books by creating control groups in different parts of the world.

What we all found so bizarre was that this kind of practice - testing, incentives, controls and so forth - was often seen by marketers as direct marketing's nerdy and embarrassing secret. And yet, in another context, it has now emerged as one of the most promising routes to end world hunger.

Duflo has discovered what most direct marketers instinctively learned quite early in their working lives - which is the disproportionate importance of changing very small things. How tiny adjustments to interface design or offers or copy may have unexpectedly large effects on human behaviour and sales. That redesigning a coupon can affect behaviour every bit as much as reworking the proposition.

In recent years, we started to scorn these things a little. Not because they were unimportant, but because they seemed small-minded, tactical, slightly awkward. When you can have a long debate about strategy, it may seem a little infra dig to spend 20 minutes debating the intricacies of coupon design.

We suppose Duflo must feel the same way sometimes: when her economist colleagues are debating the Fed's policy on quantitative easing, it must feel a bit embarrassing to butt in with: "And have you seen how well my free lentil offer is doing?"

But there's another reason why we in this industry - especially the digital community - spurned some of this great culture of testing: people became resentful of what they saw as rules. The lists seemed unduly restrictive to those of a creative disposition.

This was a misunderstanding. For they aren't rules at all. To a scientist, they would be described as findings. Or, as David Ogilvy put it when he was once accused of being dogmatic: "They aren't rules, you fools - they're tools."

What's all of this got to do with digital? We think this experimental line of activity is even more powerful in the digital world than in the analogue one. For a start, in digital-direct marketing, there are far more variables to test - and there is far more opportunity to test them. Better still, the whole process of testing and refinement can be many times faster.

But therein lies a possible problem. That the volume of information generated is so large and arrives in successive waves so fast that it never gets properly recorded or stored. Or shared. We need a mechanism by which the reams of stuff that emerge from our work every day can be stored, codified, updated and shared.

The problem is: who could possibly collect and share the kind of information that digital-direct marketing generates every day?

We thought: "What if we all did?"

What if we created a wiki to store all of our agency's learnings on everything from banner animation to check-out procedures?

Then we thought: "Why stop with one agency?"

Most of this stuff is far less valuable when proprietary than when open. It's not competitive information - and ours is not a zero-sum business. If, say, 0800 numbers work 20 per cent better than 0845 numbers, it benefits the whole industry to know this.

To start the ball rolling, we've e-mailed everyone in the OgilvyOne Worldwide network and asked each of them for a single finding with which to seed the site.

But this isn't just about OgilvyOne. It's about our industry. We want you, our peers and competitors, to contribute too. You'll find our new site here:

Feel free to amend, qualify or enhance anything you find here. All we ask is that, to use the site, you add something valuable of your own once a year.

It won't cure world poverty. But it can make our working lives a little richer.


- Digital has transformed direct marketing's nerdy secrets (testing, incentives, controls) into a fantastically powerful tool

- In digital-direct marketing, there are many more variables to test, and refinement can be much faster and more accurate

- OgilvyOne is building an open-source digital-direct wiki to act as a knowledge bank for the whole industry.

- Annette King is the chief executive and Emma de la Fosse is the creative director at OgilvyOne London