On Monday, DataXu released insights from a recent sample of client display ad campaigns that showed creative has more impact on conversions than context or audience. Read more on the DataXu blog about "Beyond Audience: What Drives Campaign Performance?" And, download the one-sheeter (PDF).
Sandro Catanzaro, VP of Products at DataXu, discussed the study's findings.
So are you saying that if context and audience remain equal (or remain the same) that nearly 1/2 the time, a change in creative led to a conversion? How do you pull out the fact that creative is the key here nearly 1/2 the time?
To provide some context, we initially conducted this analysis as an engineering effort. We wanted to see if our data would reveal any insights about which type of impression attributes are most important to a campaign. We analyzed converting impressions across a set of 19 of our campaigns for 30 days, and we found that for almost half (48%) of the campaigns, creative attributes of those impressions were more correlated to conversions than attributes related to context or the consumer. In the other half of the campaigns (52%), context or consumer attributes were more frequently correlated.
So, getting back to your question, for the 48% of campaigns in which creative was the most important attribute, with all else being equal, choosing the best creative was the most important decision driving the increased conversion rates.
While creative attributes were the "winner" in our study, the real takeaway is that there is not one approach or dimension of optimization that fits all campaigns, as campaign performance drivers can be unpredictable. Ideally, you need to consider all three data domains, at the same time—and for every impression— if you want to maximize performance.
Given the limitations of certain inventory sources and exchanges, did any of the campaigns run across non-RTB-enabled inventory? If so, do you have any estimate of the performance impact of a non-RTB-enabled environment?
The analysis in MarketPulse was completed using data from both types of exchanges, but we didn't break out the data according to RTB and non-RTB to see how that impacted results. Maybe that will be our next topic!
This issue, however, really does get at the heart of the difference between RTB (impression-by-impression) and non-RTB (static rules-based) buying. The RTB market means that advertisers who have partnered with a technology provider (like DataXu) are given the opportunity to apply finely-tuned optimization strategies to each and every impression, and adapt quickly in response to results. In a non-RTB environment, you must pre-define "segments" for your buys, with flat pricing; this forces you to use approximate optimization, which limits performance. Advertisers who are not using RTB aren't getting the insights, performance, or ROI that they could be achieving in a more dynamic buying and optimization environment.
Are clients able to take full advantage of RTB-enabled buying today? Where can they be more aggressive around optimization?
Many advertisers are not taking full advantage of RTB yet. The majority of potential customers we talk with are either not using RTB in their media mix, or they are using it for simple audience buys and retargeting only. This is leaving a lot of potential on the table, since they are not taking full advantage of RTB's capabilities.
Even simple campaigns can benefit from RTB-enabled buying. The right platform partner definitively makes it easy to manage scale and complexity (such as hundreds of creatives!). Although it may feel counterintuitive, one way that clients can be more aggressive about optimization is by shifting more of the burden to an advanced demand-side platform—one that can quickly assess campaign performance patterns and put that knowledge to work immediately. This can eliminate some guesswork at the outset or reliance on broad insights derived weeks—or even months—into a campaign to adjust a campaign's strategy.
We're not suggesting by our results that advertisers abandon their preferred optimization tactics related to consumer, context, or creative. It's really more of a recommendation to not assume too much at the outset and to also take advantage of the more comprehensive, dynamic optimization capabilities available in the marketplace today to maximize opportunities for success.
By John Ebbert