Commentary

Not The Last Word On Profile-Based Targeting

  • by , Featured Contributor, January 18, 2002
Last week’s news about DoubleClick shuttering its profile targeting business has many in the industry writing the obituary for all. The thinking is that if the Big Dog on the block can't make it work, certainly no one else can. I don't see it that way.

To me, the announcement says that if you're not a branded content publisher already selling your advertising at a premium based on explicit data about your audience, profile-based targeting may not work for you because it may not add enough value to pay for the cost of targeting. Why do I think that?

I think that DoubleClick's business model stood in their way. While I certainly cannot speak for them or their motives or have any special knowledge of their operations, I believe that given their business model it was almost impossible for them to efficiently collect, manage, target and monetize audience profiles in online advertising. I base that conclusion on the following reasons:

1. Efficiency - Profile-targeting with a network-based system is costly from a technical standpoint. As I understand it, people’s profiles in DoubleClick’s system were stored in a database on the DC network rather then in a cookie on a person's browser. That means that to target an ad, you’d have to look into a database every time that you serve a profile-targeted ad, which is a costly task. However, sites like New York Times Digital, which use cookie-based profiles, have no costly database "look-ups" and can deliver targeted ads at a significantly lower cost per ad. That aside, branded content sites like NYTimes.com regularly sell $10+ CPMs and can generate a lot more money from a 25% lift in rates for profile targeting than someone who’s selling at an average $2 or $3 CPM. Bottom line - expensive targeting and very little net revenue lift.

advertisement

advertisement

2. Data Ownership - DoubleClick could never collect the personally identifiable data or user-declared data like zip, age and gender, because they didn’t have direct relationships with their users/audience. Therefore, without zip, age and gender, DoubleClick lacked the essential information most advertisers need for true targeting.

3. Privacy - The data DoubleClick could collect from its interactions with publisher/partners was limited and very thin - while they were able to build an enormous number of these profiles, their lack of owned "opt-in" data meant that they had to add data to the profiles to really build any value in them. And, as we all remember from two years ago, privacy concerns prevented them from merging their Abacus database or any other Personally Identifiable Information databases into their online profile database. Thus, while they had a great foundation and millions of anonymous behavioral profiles, privacy concerns prevented them from building anything truly valuable.

So, DoubleClick was faced with those issues related to their profile-targeting business, but at the same time the company was having great success in their technology and data businesses. Therefore, I’m not at all surprised that they determined to exit the profile targeting business. But to interpret the exit as an industry-wide obituary for profile targeting would be very wrong. Let's not forget that it was only a month or so ago that New York Times Digital announced that their online media operations were now profitable, and that registration-driven profile targeting is the centerpiece of their advertising strategy. It works for them; and given ongoing initiatives by branded content publishers and organizations like the Online Publishers Association and the IAB, I suspect that it will work for many more.

-- Dave Morgan is the founder and former CEO of Real Media. He is currently CEO of Tacoda Systems, formerly known as True Audience – an audience management software company.

Next story loading loading..