BigPond/Telstra: Who Better To Tackle The Impossible?

In the over-exposed setting of Cannes this week, Justin Milne, BigPond’s CEO, announced a mobile advertising strategy that is certain to have consumers cramming the exits looking for the nearest Optus reseller

His goal is to offer the Australian marketplace the most sophisticated, most complex system of ad targeting there is anywhere in the world and will use Telstra’s 3G network to deliver the capability.

His behavioural targeting utopia is to triangulate the individual’s location, mobile and online data, then overlay that with some trusty customer segmentation of its own, to arrive at some quantitative assumption (read: scoring) about a person’s purchase intentions.

Based on this theory, an individual with a Telstra 3G handset, who’s walking down Norton St in the Sydney suburb of Leichhardt, for example, would receive a Yellow Pages ad for a local restaurant, perhaps with a discount offer to sweeten the deal.

That ad could be based on any number of variables, including the fact that the person was in the general vicinity of the restaurant, or they had previously searched for “restaurants” on their mobile browser, or had browsed a website like CitySearch, or from a socio-demographic perspective matched an ‘urban sophisticate’ profile. Of course, the ad could also be triggered by any combination of these four elements.

The question will be how each ‘input’ is weighted to ensure the system has optimised to reach a person with a high enough propensity to buy a meal at the restaurant. Without the ‘smarts’ to do this, the likelihood of the person taking up the offer will be relatively low, and the restaurant will have waste significant advertising dollars on a very expensive targeting system (relatively speaking).

Milne’s ambition is admirable. His system is world-class (in a theoretical way), but there are some obvious holes in the data architecture sitting behind the system, not the least being how to capture online data to fit into the subscriber’s profile.

How does BigPond get its hands on data generated from non-Telstra related websites? If I’m a frequent news.com.au reader, with a penchant for restaurant reviews, how does BigPond know this, and if it doesn’t have access to this type of third-party data (online or off-deck in a mobile context), isn’t this already undermining the integrity of the ad targeting system, and therefore reducing its value to advertisers?

At best, BigPond is likely to have access to very limited subscriber data, as well as stats re: on-deck behaviour, perhaps data from third-party, off-deck sites (with permission), and perhaps (and it’s a big perhaps) data from Telstra-related sites (like White Pages, Citysearch, Whereis etc. And we’ll conveniently ignore the issue of cookie deletion). The only reliable data point in the whole system will be a geographic fix on the owner’s handset.

As for customer segmentation methods (like Roy Morgan, Mosaic or some other proprietary system), these are simply too broad and too generic to add much value by way of specific ad targeting.

This is the trick. A number of consumer research pieces indicate that individuals are partial to advertising delivered either online or via the mobile handset so long as it meets their criteria for relevancy, ie, it matches an actual need, either immediately or over the medium-term.

If BigPond gets this wrong, and the advertising becomes both frequent and inaccurate, then this so-called media player will be in a whole world of pain.