Find the Right Keywords with Data-Driven Attribution

Jannette Flores / December 2018

The customer journey is more complex than ever before—customers often switch between online channels before making a purchase. This also affects our assessment of the various touchpoints. Attribution models, such as "last-click attribution," have long been considered obsolete. Instead, advertisers gain a much more accurate picture of the customer journey when they use holistic models, such as data-driven attribution, uncovering additional potential in areas that previously received too little attention (e.g., generic keywords or mobile clicks).

What is data-driven attribution and why should you move away from last click attribution?

Data-driven attribution offers many advantages over rule-based attribution models, as it provides a much more complete picture of buying behavior. Unlike rule-based models, data-driven attribution is a dynamic model that individually calculates the value contribution of each touchpoint along the customer journey. In addition, it also considers cross-device conversions, i.e. conversions initiated on one device or browser and completed on a different device or browser. Another advantage is that the data-driven model is adapted dynamically in real time.


"The limitations of last-click attribution made it difficult to justify significant investments in keywords upstream of the final click. This has changed with data-driven attribution. I have known about and used Google Ads from the very beginning—this is definitely one of the biggest improvements!"

-Tim Dunker, CEO of Berge & Meer.


Data-driven attribution: How to find matching keywords

Attribution models also give you insights into the contribution of each keyword. The problem is that many generic keywords are more likely to be effective at the beginning of the customer journey, that is to say, when customers first look for a particular product or service. As a result, marketing managers often classify them as "too expensive" because they generate little or no conversions in a last-click model. In this case, data-driven attribution models can outperform the rigid models. With data-driven attribution, the value of upstream search terms can be depicted more realistically and, for the first time, is dynamically adjusted. This helps advertisers understand how to successfully grow their keyword portfolio to reach customers earlier in their decision-making process. The travel agency Berge & Meer, for example, benefitted from switching to a data-driven attribution model and was able to find suitable keywords.

The following tips can also help advertisers optimize their keywords:

● Identify the right keywords: The Keyword Planner can be used to search for the most relevant keywords for a product or service. In addition, advertisers receive suggestions on bid estimates and learn how many times users have searched for their keywords.

● Go to the search terms report: The search terms report tells advertisers what performance their ads have achieved when served as a result of searches on the search network.

● Re-evaluate keywords: Give paused keywords that have not performed under a last-click attribution in the past a new chance to be reassessed with data-driven attribution as a whole.

Here's how it works: apo-rot boosts conversions with data-driven attribution and keyword extension

Thanks to the transition from a "last-click model" to a data-driven attribution model and keyword extension, the apo-rot online pharmacy has been able to better take into account the actual value added in its campaign management. The change brought noticeable effects and increased efficiency. Shopping conversions increased by 24 percent and mobile shopping conversions by as much as 40 percent.


"Customers come through different channels. Our goal is to better understand the customer journey of our customers. We want to know when we need to connect with users to invest our budgets more efficiently."

- Sarah Küter, Online Marketing Manager at apo-rot


A second optimization approach was to revise the set of keywords. Data-driven attribution offered a new way to test keywords that had previously been paused due to low target values when ​​viewed from a last-click perspective. In conjunction with the internal SEO team, many generic keywords were added, such as "ginger capsules," which play a role in the early phase of the purchasing process. 3,000 new active keywords increased campaign profitability. As a result, the conversion rate increased by 6 percent and the click-through rate by 9 percent.

The results from the attribution report strengthened the importance of mobile. Mobile results showed a conversion uplift of 26 percent and a 20 percent lower cost per acquisition (CPA). By capturing cross-device user behavior, a budget shift towards mobile could be even more justified.


"The acquisition of new customers is at the top of our agenda. With the new upper-funnel keywords, we're reaching significantly more new users today."

- Sarah Küter, Online Marketing Manager at apo-rot


Outlook: Google Attribution to launch in Germany

In the future, Google Attribution Beta will enable advertisers to optimize the impact of their digital marketing activities across devices and channels.

How to unlock the power of first-party data