Attribution fraud is typically encountered within the context of digital marketing. It occurs when publishers intentionally or unintentionally try to “game the system”— specifically, the marketer’s measurement system — by taking credit for conversions attributed to their media.
Who is at Risk?
Marketers that measure their campaigns based on last touch, first touch or other rules-based attribution methodologies are the most prone to be the victims of attribution fraud. These methodologies take a basic rule for attributing conversion credit, and arbitrarily apply it to all channels, publishers and tactics in the marketing plan. As a result, some publishers will strategically serve their ads in a way that enables them to take 100% credit for a conversion, regardless of whether or not their ads actually influenced the user to convert.
For example, when marketers use a last-touch methodology, some publishers will use retargeting tactics in placements that are not intended for retargeting. This is an attempt to win the last touch right before a likely conversion, so they can take 100% credit for that conversion.
Not only is this practice a misuse of the marketer’s budget, but it’s also not transparent to the marketer. As a result, digital media buyers are misled to optimize budget toward publishers gaming the measurement system, while potentially taking budget away from publishers actually influencing conversions.
Even marketers that use an algorithmic attribution methodology can fall victim to attribution fraud if their algorithms are not robust enough. Weak algorithms that only model at a high level, such as channel and campaign, and then cascade results down to granular levels, such as publishers and placements, are also easily gamed.
Additionally, these algorithms often only use data from the marketer’s pool of converters, completely ignoring all of the insights that come from also looking at non-converters to understand what’s working and what’s not.
How to Prevent It
There are three things marketers must adopt to effectively prevent attribution fraud:
Account for converters and non-converters in their attribution modeling.
Select the right metric for measurement and optimization.
Use a robust algorithmic attribution methodology that integrates top-down and bottom-up models.
We have already discussed why it’s important to account for converters and non-converters in the attribution model.
But selecting the right metric for measurement and optimization is also critical. This is very similar to the precaution against ad fraud. It is important to select metrics like revenue or lifetime value as the overall marketing KPI, rather than strictly using visit-based metrics like leads or page visits, which may be more easily tracked by publishers to detect patterns for attribution fraud.
Finally, it’s important to remember that the conversions, revenue and lifetime value driven by digital marketing channels may be a small portion of the total value of all marketing. An integrated top-down and bottom-up modeling approach accounts for the incremental value of all marketing channels to provide a truly holistic view of what’s working and what’s not.