The co-founder of one of Europe's most advanced behavioural targeting firms gives an honest, personable account of the struggles faced by the industry to gain the respect of advertisers.

Once again, targeting will be a central theme of this year's ad:tech London event and for good reason, in my opinion. There are many indications that targeting will turn the advertising market on its head within the next decade, and not only in terms of internet advertising!

Targeting has been a mega-trend for around three years now, but it is generally unknown that predictive targeting has existed for a good deal longer -- its roots reach back for at least 20 years and one of the most important of these roots lies in Germany.

Targeting was grabbing attention back in 2006 in Germany. Test projects were run, the air was filled with press releases and it was raining financing rounds. The second phase of hype around targeting, previously known as personalisation, had begun. At first, we at nugg.ad thought that we would now simply need to lay out our wares and demonstrate our know-how in a system that that had been fully redeveloped from the foundation up, but it did not prove to be quite that easy in practice. Online advertising revolutions do not happen overnight. I remember well one initial meeting we had with Jochen Rabe from Interactive Media. He asked us how we intended to address the ad server and we had no idea. We knew a little about ad servers, which had been interfaces in previous projects although mostly in the form of a complex back-end integration, but how to address a third-party ad server was only clear to us in theory. Other problems cropped up later, for example, forecasting for targeting campaigns, guaranteed bookings, reporting and so on. We learned the hard way that although everyone was talking about targeting, in terms of day-to-day bookings no one had any experience (ourselves included). We were setting off on a long journey.

For a long time, we can freely admit today, we were not completely sure that everything would work out as we had planned it, i.e., being able to predict a target group based on online click behaviour combined with market research data. As mentioned above, the idea itself was already fairly old and in some ways also tried and tested. Even the algorithms being used were well known to the scientific community, but whether it would actually work, that it would result in better campaigns with lower media loss, remained a vision for a long time. We were taking a big enough chance as it was, not to mention all of the (indirectly related) unanswered questions such as the business model.

In our initial enthusiasm and naiveté we initially proposed a true revenue share. We then learned very quickly that although everyone was very taken with the idea, no one wanted to make the turnover being generated by targeting campaigns public knowledge. We also had to learn that in spite of all of our arguments about the increased efficiency produced by targeting being based on scientific facts, these were often simply undercut by discounts at the point of sale.

Neither did the complexity of the subject decrease. After a short time, once we were able to find our way around the various types of ad server, we suddenly discovered that ad networks could create totally new problems for us. All of a sudden, it was no longer possible just to place an advert on a site for a single day, because the ad-network did not own the space. In addition, we were no longer looking at just two or three portals, but suddenly upwards of a thousand websites, which were not reserved exclusively for our advertising campaigns. How was targeting supposed to work under these conditions? And how were we supposed to treat the content in the absence of any categorisation, such as the AGOF-/InfONline? We are now two years further down the road now and have made a lot of progress. The technology is tried and tested and its effectiveness has been proven, even in ad-networks, and we even have a business model that is acceptable to everyone.

It will continue to require a great deal of training, education and persuasion, but the targeting revolution is coming. I am more convinced of this today than ever before. In fact, we are right in the middle of it! But it will be hard work, it won't be easy and it won't be quick. The technology still has to be improved and we will probably still be fighting the battle to convince people of its merits in two years' time. This is why I am already looking forward to ad:tech London 2010 -- where targeting will continue be a hot topic, once again!  

Stephan Noller is managing director and founder, Nugg.ad.