Online behavioural targeting offers - in theory at least - a way of predicting what a consumer is likely to do next. So it's easy to see why marketers are getting a bit excited about it. Its central premise is that past behaviour is a predictor of future behaviour. It is simple, yet potentially radical.
Early results have been impressive. Tiscali UK began allowing advertisers on its 25 channels to use behavioural targeting in their campaigns in early 2007. Alex Hole, online media director at Tiscali UK, says: "We've seen click-through rates on many of those campaigns increase by up to 500 per cent. What is more, our users are now getting served with ads that are more relevant to them."
There is still a great deal of confusion and scepticism surrounding behavioural targeting. Not everyone believes it can work, few marketers know how to make it function for large-scale online campaigns, and even fewer have actually tried it. Yet, the potential is so great that no one can afford to ignore it.
So why are marketers slow to jump on this particular bandwagon? A significant factor in the slow take-up of behavioural targeting is the general confusion about what it actually is. Jim Sterne, chairman of the Web Analytics Association, says: "Very often, people mistake contextual advertising for behavioural targeting. Contextual advertising is easy to understand because we've been doing it in print for ages. If you sell golf clubs, you'll want your ad to show up on a website all about golfing."
TMP Worldwide recently ran a campaign for GCHQ, a government intelligence organisation. TMP placed billboard ads inside six Xbox Live games to encourage recent graduates with an aptitude for problem-solving to visit the GCHQ recruitment website. Andrew Wilkinson, chief executive of TMP Worldwide, says: "The campaign was highly effective. Three thousand people visited a specific landing page advertised in the game, and it produced a great uplift in visits to the recruitment website."
But while this may have been a successful campaign, it was not behavioural targeting in the purest sense. It was targeting by context. Sterne goes on to define behavioural targeting: "This is showing your ad for golf vacations to people who have exhibited a specific interest in the topic over time. It's more than just reading one article about one faraway golf course, but consistently choosing articles, typing in searches and clicking on links that indicate a propensity toward your offer."
There are two ways of applying this principle online. First, you can use business rules. For example, if a user clicks on a certain number of articles that are tagged with the appropriate keywords, then they get an ad associated with the subject matter. Second, you can use collaborative filtering, which is where you predict the ads a consumer will want to see, based on their past behaviour and the stated preferences of other people who have exhibited the same behaviour.
For most, however, the greatest obstacles to running this type of campaign are practical. Louis Fernandes, account director at online marketing agency Acxiom Digital, says: "There have been very few examples of online ad campaigns using behavioural targeting. Generally, this is down to how organisations plan to use data in the context of the campaigns that they run, how they collect this data, how they process it and finally how they store it. The sheer volume of data available is outstripping the processing capabilities available to most organisations."
Ray Welsh, sales and marketing director at digital marketing agency Mailtrack, agrees that finding meaning in a large volume of click-stream data and altering the next communication based on that knowledge is very daunting for many marketers.
However, he says: "This process isn't actually half as complex and expensive as most people think it is. The CRM functionality needed to gather, store and analyse behavioural information is no longer only found in high-end systems."
Yet, marketers are not rushing to try it. Kit Desai, country manager of online advertising provider Adtech, part of AOL, says: "In my experience, usage is small and selective and not large scale. Clients are trialling it and slowly building traction."
The evidence from Tiscali UK certainly seems to suggest that it can work. We can surely expect more and agencies to start picking up on this and exploring how clients can benefit from behavioural targeting. They will be encouraged by the arrival on the market of more software that automates the tricky data collection and processing. We can also expect greater integration of offline and online behavioural targeting data to produce even more accurate and effective customer profiling.
As Mike James, managing director at digital advertising network Adconion, concludes: "Behavioural targeting is fast coming of age. I expect that as large brands start to dip their toes in the water, more and more will start seeing the benefits of creating sophisticated messaging tailored to an audience that is becoming increasingly sophisticated in its buying habits."
HEALTHIER WEIGHT CUTS WASTE WITH BEHAVIOURAL TARGETING
Healthier Weight is a Birmingham-based company that offers surgery programmes to the morbidly obese. It offers gastric banding and similar operations by surgeons in Italy, as well as a range of food and drink products that are designed to reduce weight.
It hired online marketing agency Acceleris Marketing Communications in 2006 to drive traffic to its website and then encourage visitors to the website to make a phone call to the company. Acceleris began with small-scale campaigns placing ads on health-focused websites, such as netdoctor.co.uk.
Jane Slimming, digital specialist at Acceleris, says: "This just wasn't getting the volume necessary. To meet its sales targets, Healthier Weight needed to increase the number of visitors to its site, and it needed more of those visitors to phone the company. So the team there came to us and asked us to run a larger scale campaign."
Although the budget for this campaign was increased to £50,000, it was still necessary to make this work as hard as possible. So Slimming went to Yahoo!'s blind network. She explains: "Blind networks are a great way of getting a large number of impressions for low investment. They're much cheaper because you don't know which sites they're coming from - they're blind - and so you tend to get a lot more wastage."
However, by using behavioural targeting, Slimming was able to cut this wastage. She says: "Yahoo! has just started offering behavioural targeting on its blind network and it's great. Basically, it only serves our ads to those users who have behaved as though they're interested in dieting. Those behaviours might be clicking on a diet ad or entering a dieting term into a search engine."
The time users spent on the website also increased, indicating that the campaign was indeed attracting people with a genuine interest in Healthier Weight's products and services.
Slimming is convinced behavioural targeting is appropriate for large-scale campaigns such as this. She says: "Previously, the ability to target was limited to matching up the message to the site. For example, if selling cars, you'd advertise on a niche car-enthusiast's website. This type of targeting saw high click-through rates but couldn't deliver a broad reach."
She concludes: "Demographic and geographical targeting are steps in the right direction, but are limited in their recency and accuracy. It's only behavioural targeting that allows you to reach a large number of people who have recently done something to suggest that they are currently interested in what you have to offer."
Comments