Marketers are finally beginning to embrace multi-touch attribution (MTA) as a way to make in-flight optimization decisions in market, because they can learn which channels, platforms, publishers, formats, creatives, and audiences are influencing purchases, and how they all work together. MTA represents a major step forward for accountable marketing when it’s well understood, implemented correctly, and used in combination with other measurement methodologies like marketing mix modeling (MMM).
In fact, CMOs across the world agree that customer experience, customer retention and growth, and customer analytics are crucial to support marketing outcomes over the next 18 months, according to Gartner. It is, therefore, no surprise that a study by the Data and Marketing Association (DMA) and the Winterberry Group concluded that two-thirds of U.S. marketing professionals had increased their prioritization of marketing attribution in the past year.
But as valuable as MTA can be for marketers looking to improve their ROMI, many first-generation attribution solutions and general confusion in the marketplace have caused a series of myths to cloud the promise.
Here are six commonly misunderstood assertions about multi-touch attribution that every marketer should know about:
MTA fails to measure incremental impact
Some old school mix modelers have suggested that MTA is flawed because it is limited in scope and can’t take into account many of the key drivers influencing sales. They claim that MTA over-attributes marketing success to the channels/devices it reads and under-attributes the impact of everything else.
While that may be true for solutions that are limited in scope and explanatory power, best-in-class marketing analytics offerings take a unified approach, which brings together traditional non-addressable media, addressable media, and non-marketing drivers such as weather, economy, seasonality, competitor activity, etc.
This approach lets marketers measure the incremental impact of media. So not just the interaction between addressable channels, such as display, paid search, paid social but also channels like Out-of-Home (OOH), radio, print, TV, etc. If all advertisers look at is digital/addressable, at best all they can understand is their relative influence on each other. When considering “everything,” media, as well as non-media business drivers, advertisers can accurately determine the incremental impact that media has on their business. Marketers really need to think of this as business driver modeling and not just marketing mix analysis.
MTA fails at transparency and objectivity
OK, this is fair with some media platforms that also provide their own measurement (and have a tendency to over-attribute their own influence and under-attribute other publishers). But for the most part, advertisers have objected to these offerings and demanded neutral measurement providers. Marketers want and need an objective view into their performance and value the ability to look across all customer touchpoints.
They demand neutral, trusted partners they can work with to ensure they are getting an unbiased view of their media and are able to scale their advertising programs beyond one digital platform provider. With neutrality comes transparency and objectivity.
MTA doesn’t measure at the customer level
The promise of MTA is understanding the impact of media at the customer level. Without this degree of granularity, MTA doesn’t work. While customer-level measurement is challenging – it is difficult to reconcile customer identity across devices, channels, platforms, browsers, etc. – some advanced measurement solutions available in the market today are able to join these disparate data sources into a single individual or household level view. It’s important to probe any potential provider about their view of identity and the scope of their graph. Strong MTA requires a broader aperture than merely a cross-device or cross-channel view.
MTA has to be real-time to make a difference
The fallacy of real-time MTA is pernicious. The vision of real-time media allocation is attractive, but unrealistic. While there are certainly moments when it makes sense to quickly move money in or out of the market based on external factors, most advertising works on a slow build, cumulative basis. And even if a single ad were all it took for a consumer to buy; the purchase cycle is usually less than immediate.
Worse, real-time attribution solutions are by definition last-touch, which is the antithesis of how advertising works (see above). By being able to better understand the whole journey, marketers will inevitably receive better outcomes, in the form of lower CPAs, higher ROI, etc.
Therefore, MTA can’t inform the media buy
Based on the argument above, the best MTA solution is not last-click and takes time to read. True. But for MTA to be truly valuable — even though decisioning isn’t real-time off of media exposure — marketers do expect to be able to make in-flight optimizations. By measuring both media attributes (e.g., campaign, site, placement, creative), as well as audience attributes, advertisers can use them both as direct inputs into the buying process. For example, advertisers can identify an audience that works well in their MTA results and create a look-alike model for that audience segment. They can then push that segment into their favorite programmatic platform. Obviously, this requires that the MTA platform be connected directly with the media platform so that this customer analytics feeds customer experience flywheel can be successful.
MTA is static and hasn’t evolved as new forms of data emerged
New forms of data and new rates of granularity have made previously opaque patterns of consumption and behavior transparent. Data lakes allow unstructured data to be used for analytics, and MTA can use long-term longitudinal data but isn’t hostage to them. We’ve seen a lot of progress, especially with the leading digital platforms like Amazon, Facebook, Google, Snap and Pinterest – they’ve opened up access to their data sets so advertisers can get an accurate, holistic view. The best providers are always looking at new data they can address and how much of the media mix can be included in the model.
MTA isn’t the holy grail, to be sure. And bad MTA is worse than no MTA at all. But with the right provider, a strong and holistic identity graph, integration with MMM, and a direct connection back to your media platform, MTA provides actionable analytics and intelligence that strengthens marketing returns for all consumers, not just those already down the funnel. Don’t let the fallacies get in the way of your strategy. MTA myths have already been busted.
This article was originally published on MarTechSeries.com on July 31, 2018.