In recent years, Google’s been highly invested in educating advertisers on various attribution models, and making data sets readily available and easy to analyze.
The old way of relying on last-click attribution was often problematic and limited. It essentially ignores early touch points that sow seeds for future conversion harvests, making it tough for marketers to optimize budgets to enlarge the best channels for finding new customers. Branded search always did great. How could it not? However, other display and non-branded efforts were often erroneously cut back, with less provable value than they deserved.
This year, attribution model choices may be getting slimmer, or perhaps a new default will change our habitual way of counting our conversions. Although no public statement from Google, many SEMs have claimed or are beginning to wonder if the relatively new (and highly recommended by them) “data-driven attribution” method will become the default, replacing the decades-old “last-touch attribution” model. Whether or not this happens (and when), now is the time to study your models and proactively make a choice.
Serious advertisers have used alternate models for years now. When the strategy is for aggressive new-customer growth, then the “finding” channels (top of funnel) become more important. Consider a first touch, position-based or custom modeling to keep the credit (and the budget) going to the top-of-funnel channels, where it belongs.
If it’s important to give every touchpoint channel credit, then a “linear” attribution model is perfect, dividing up all touches and giving equal credit. Should the newer marketing efforts trump the older campaigns from months ago? The time decay model is for you. With six pre-packaged models to choose from, and room for customization, lots of options are on the table. Pick one that makes the most sense for you, and adjust your targets and expectations accordingly. Many products make it easy these days, and Google Attribution’s data-driven attribution capabilities are available for those that want a good look under the hood.
Will attribution data ever be 100% perfect? Probably not. And that’s okay. The idea here is to do something. Reassess your digital ad strategy to consider the bigger, holistic picture. Gone are the days when savvy digital teams consider their own channels in silos. Cross-channel consumer experience should be improved at all costs, and one of the best ways to understand the ebb and flow between channels is by studying these models.
In my 2018 paid search strategies post, reassessing attribution made the narrow list. Whether or not the long-standing last-click model stays default, a fresh look at what makes most sense for your business is never a bad idea. With so many new features, platforms and technologies, there’s little reason to be content with a 15-year old default method for reporting some of your most important data.
