With so many different analytics systems, Collective believes that causal attribution will have a tremendous and positive impact on the digital advertising industry. Here's why.
At a time in our industry when everyone is struggling with analytics and the increasingly dire need to measure display advertising's effectiveness, Jeremy Stanley, SVP of data and product services for Collective, made the clear and concise case during a Spotlight presentation at the iMedia Brand Summit that causal attribution is the solution we've all been waiting for.
"All metrics are misleading at best," Stanley said. "At worst, existing metrics destroy value. So what do we do? How do we tackle these challenges?"
Based on the assumption that the current methods of measurement are deeply flawed by a reliance on "weighted or algorithmic" attribution, Collective's point is that billions of dollars are being wasted on industry standards that no longer apply, are too-easily manipulated, are too complex, too subjective, and continue to do a disservice to the advertising industry as a whole.
Causal attribution, however, directly quantifies ROI by measuring outcomes caused by advertising, Stanley said. It is also unbiased by all other advertising, is transparent to administer, provides rich audience analytics, and it uses rigorous experiments to measure the increase in desired outcomes caused by display advertising, verses a correlation effect measured by existing attribution models.
According to research conducted by Collective, current online attribution models are flawed in the following ways:
Algorithmic attribution: A computer algorithm assigns credit for outcomes by analyzing data. This method can sometimes overcome limitations in simpler metrics, but it confuses correlation for causation, requires measuring all user interactions in all channels, there is no industry standard, and only programmers really know how it works.
Click-through rate: Measures clicks per impression, but very few people click, clickers rarely buy, most clicks are accidental or fraudulent, and the system is easily gamed by placing ads near high-click content, such as games.
Last-click attribution: Measures the last click prior to outcome. This system only measures "productive" clicks and is hard for intermediaries to game, but the ads that build awareness or intent are given no credit, and many users would have converted anyway.
Last impression: Measures the last ad prior to outcome, but the ads do not have to influence consumers to receive 100 percent credit, the quality of media, placement, and creative have little impact, it is easily gamed, and it overvalues retargeting.
The advantages of causal attribution as an alternative, Stanley said, enable advertisers to "accurately evaluate how their online advertising affects the behavior of specific audiences and measure the value generated by their advertising spend."
Among other advantages, some of the key findings Collective presented on why causal attribution can be an advertisers best friend included:
- Causal attribution provides an unbiased method of linking true ROI to advertising spend, including exposing campaigns that aren't working.
- Because it is cookie based, ROI can be cascaded down to individual audience segments, providing rich insights into the types of users who are being influenced by the advertising.
- Causal attribution can be used to measure offline conversions and brand lift.
- It directly compares the performance of different sources of media
- It tests multiple, potentially radically different creatives to identify which audiences are influenced by each.
- It quantifies the effect of frequency -- how many times each cookie is shown an advertisement.
- Proper ROI measurement provided by causal attribution leads to increased ad spending, higher returns for advertisers, and an incentive for publishers to create quality and engaging content.
Gretchen Hyman is editor-in-chief of iMedia Connection.
On Twitter? Follow iMedia Connection at @iMediaTweet.