Study Shows Behavioral Targeting's Strengths, Weaknesses

A new behavioral targeting research report from 24/7 Real Media sheds light on the subtleties of the emerging interactive marketing practice.

Data from three fourth-quarter case studies show behavioral targeting can boost reach and return on investment (ROI), or lower cost per acquisition (CPA) in some cases. But the research reveals it’s ineffective in others.

“We found that we could not make sweeping generalizations,” said Jack Smith, VP of product strategy at 24/7 Real Media. “We confirmed a lot of things we expected, but we want to see how it changes over time.”

Campaigns using 24/7 OnTarget, which launched in March 2004, split visitors within the 24/7 Web Alliance network into eight behavioral segments — Entrepreneurs, Money Makers, Music Lovers, Techies, Globetrotters (travel), The Better Half (women), Health Specific and Gamers.

Starwood Hotels and Resorts used the Globetrotter segment for a campaign to drive U.S. travelers to its Latin American resorts site. In the 45-day campaign, it found behavioral targeting helped lower customer acquisition costs and boost ROI by 22 percent, when compared to a single-site sponsorship of a major market newspaper site’s Travel and Hotel sections.

In another campaign, this one for a large Web hosting company, behavioral targeting let the advertiser extend its reach to targets that were not available through contextual targeting, while achieving the same or better overall ROI. The 30-day campaign expanded an existing site sponsorship of a technology-oriented Web site to include behaviorally targeted ads on 24/7’s network. Segments targeted were Entrepreneurs, Money Makers, and Techies.

The behaviorally targeted ads reduced customer acquisition costs for the Web host’s managed services by 61 percent compared to the site sponsorship. In addition, clickthrough rate (CTR) doubled while conversion rate was similar to the more expensive site sponsorship. As a result, cost per acquisition (CPA) dropped by nearly 70 percent, while the campaign’s reach was simultaneously extended.

The results were not all rosy, however. In the case of a global DSL provider, the report found that CTR for behaviorally targeted segments was low, and CPA rose dramatically. Since there was no specific segment defined for Internet users without a broadband connection, the advertiser chose to target Globetrotters, Money Makers, and Techies. While both the Starwood campaign and the Web host’s campaign targeted only those with significant interaction within the previous 30 days, the DSL campaign used a much broader approach, targeting users that had visited less frequently or been less active.

The report suggested that the poor results of the campaign may have resulted from any of several variables. One issue was the use of “stale” creative, which had previously been seen by viewers in a run-of-network campaign. In addition, the report questioned the targeting choices, noting that the segments targeted were the most likely to be early adopters of broadband, and so were the least likely viewers to be interested in the offer.

In cases like this where the target audience is not well defined by a behavioral segment, the report conceded that behavioral targeting may not be appropriate.

When researchers examined data across 1000 campaigns, they found that some behaviorally targeted ads can outperform ads that are contextually targeted by a content category. For example, in the Money Makers behavioral segment, the report found that performance in the Finance content channel was just slightly above average, while performance in the Travel content channel was twice as high.

In another case, the study found that performance of the Gamers behavioral segment in the Gaming content channel actually under-performed that of other content channels. In fact, ads targeting the Gamers segment in the Health content channel were more than eight times as effective — which the study attributes to ad clutter in the Gaming content channel.

“When we looked at our gaming content across our network, we saw a high amount of clutter, and a low share of voice among individual gaming companies. There are a lot of ads, almost all of them targeting gamers,” Smith said. “When you remove an ad from that context, it really stands out a lot more. When you’re trying to reach gamers, you may be better off marketing to a gaming segment within another channel.”