Live

Watch CBSN Live

Facebook and the Behavioral Marketing Scam

The LA Times appears to be shocked, shocked that Facebook planned to use personal data in delivering ads to users. Its reporter must have missed the terms of service and Facebook's help page on ad targeting options.

In fact, Facebook has the reputation of being the über personal data impresario. You'd think with all that information, the site would work tremendously well for advertisers. And yet, that doesn't seem to be the case, as Mike Shields at AdWeek has pointed out. The contradiction between having massive personal data and low advertising effectiveness is true in a lot of online marketing. Granted, many ads may just suck and so work badly. But maybe there's also something wrong with online behavioral marketing as it's done.

Behavioral marketing falls flat
The theory of behavioral marketing is that you get data on what people do online, learn what actually interests them, and then present ads to those who will be most receptive. The concept passes the common sense smell test. Then again, at one time so did the notion that the world was flat.

Results often tell a different story. Facebook ads don't perform well. The company should be king of behavioral marketing, given the massive amounts of information users reveal to them. (Facebook cost-per-click prices jumped by 40 percent last quarter.) And a recent survey showed 58 percent of people saying that, in general, banner ads aren't relevant to them, even with the wealth of targeting data that many outlets claim to have.

That's not to say that online advertising is ineffective. I recently spoke with the digital marketing executive for a global electronics brand. He said that in his experience, standard Flash ad click-through rates -- the percentage of time that someone seeing an ad would click on it -- were in the range of 0.7 percent. (Many direct marketing campaigns would find that level of response to be disappointing.)

Move to video ads, and he said that number can jump by a factor of 10. But if having the data and appropriately targeting ads was such a powerful concept, why would advertising on Facebook do so badly?

It strains credulity that the problem is poor concept and execution on the vast majority of ads. If behavioral marketing really were so powerful, wouldn't you expect companies to get better results? And that would lead to greater demand and, therefore, higher ad rates -- an issue of supply and demand that would trump the problem of infinite inventory the online ad industry sees.

Advertisers can't see the customers
The ad industry has sold the concept of behavioral marketing as a way to make ads more relevant to consumers and, as a result, more effective. However, something is wrong. Behavioral ads often don't deliver what the ad vendors promise. Here's why that may be:

  • Behavioral marketing assumes that there is a one-to-one relationship between people and computers. Put multiple people on a home computer, and results go haywire.
  • Automated systems can easily misunderstand why people do what they do and, as a result, assume all manner of interests and needs that don't exist.
  • There is no direct learning connection between the advertiser and the process of targeting.
For my former-direct-marketing-guy money, the third reason is the biggest. You can use personal data to better profile customers and prospects. However, implicit in every such campaign is a feedback loop:
  1. Analyze the product or service and create a best guess of the ideal target market.
  2. Find communications channels that seem to fit the profile.
  3. Build offers designed to appeal to the prospects.
  4. Communicate with the prospects.
  5. Analyze results.
  6. Refine the customer definition and measure the effectiveness of the channels.
  7. Go back to step 1 and try again.
Direct marketing is iterative. What you learn in each round helps you better understand how to reach prospects and which are likely to be most receptive. Companies deal with list vendors, agencies, and publishers, but they also get the raw data of what does and does not work. You examine how a given mailing list, magazine, or television channel performs, correlating that data with different offers and target market definitions. Campaigns improve results over time.

When a Facebook, an Apple, an ad network, or some other third party interposes itself into the process, it generally denies companies access to that raw data, because advertisers buy categories of user characteristics. They don't necessarily know the context of the ad, or even on what sites it will appear. They can only assume that the ad network is accurate in user categorizations. The process actually resembles a television buy, in which you choose programs based on demographics that might match your target customer base.

Some behavioral marketing outlets may pass on the information necessary for incremental improvement. But many, like Facebook, are black holes. That suggests one of the biggest systemic ad sales scams that the industry has seen -- and that's saying a lot.

Related:

Image: morgueFile user hogg, site standard license.
View CBS News In