Two California based companies, Adisn and Tumri seem to have captured the imagination of mainstream advertisers with their technologies to dynamically test and improve ad performance.
As reported by The New York Times, the companies’ technologies not only use behavioral, contextual or demographic targeting to work out who is looking at an ad, but also dynamically change elements of the ad to improve click-throughs, sales and other desired responses.
In other words, Adisn and Tumri have created technologies to automate the processes of (a) understanding the target audience and (b) tracking, testing and improving ads to elicit the desired response from that target audience.
Adisn’s approach is to build a database of related words in order to assess the content of a website based on the words on its pages. Adisn then buys space on a website and, in real-time, uses its ‘understanding’ of that website to display the most appropriate ad – complete with text, images and colors, etc.
Tumri, on the other hand, doesn’t buy ad space. It creates an ad template, then uses whatever targeting approach its clients are already using – behavioral, contextual or demographic, etc - to accordingly assemble an ad in real time.
In both cases, the elements used in an ad – the headline, color scheme, images, etc – are combined on the fly. Then, as the response to the ad is measured, further combinations are tested in order to improve the performance of the ad.
Real-time tracking, testing and improving is like a dream come true for direct response advertisers. But, as The New York Times reports, it poses a big challenge to the way traditional ad agency creatives approach their craft.
Such creatives – and I speak from experience, having begun my career as a copywriter for a mainstream ad agency – tend to create ads on a more holistic basis. Indeed, most aim to create ONE ad based on ONE BIG IDEA i.e. the one execution they believe will have the optimum impact.
Traditional ad creatives tend not to consider ads in terms of their individual parts i.e. that one color could or should be exchanged for another, or that one headline could be exchanged for another, or that one image could be exchanged for another.
And they have a point. The risk of approaching and tweaking an ad on a piecemeal basis is that you fail to take into account how each piece works together to generate an overall effect.
As direct response marketers who manually change and test things have found, however, I don’t think this is an insurmountable problem.
There’s still room for good old fashioned common sense – you can still avoid inappropriate combinations e.g. white text on a white background. Moreover, the results of each iteration of a given ad will provide plenty of feedback. And testing, tracking and improving sure beats not having a clue about what improves results!
Let’s face it, the days of the one ideal ad… the one big idea… and creating ads without tracking their performance… are numbered.
Source: Stephanie Clifford, “Web Marketing That Hopes to Learn What Attracts a Click,” The New York Times, December 3, 2008