Earlier this year, Google introduced Search Funnels Attribution Modeling Tool, a tool for AdWords users. According to a blog post from the company on Google+, the new tool is designed to help advertisers better understand a customer's path to purchase and to better measure the value of AdWords advertising by assigning value to the keywords, ad groups, and campaigns that lead to conversions.
It's great that Google has recognized that its current conversion tracking is antiquated, and its expanded attribution capabilities can help some advertisers better optimize search spend. But while some attribution is better than no attribution, there are definite shortcomings to Google's approach.
The tool only looks at search
Since Google's attribution modeling tool is designed for AdWords, it only measures search and neglects to incorporate data from other channels, such as display, affiliate, retargeting, social, and more. While search is an important channel for many in the advertising ecosystem, it is often not the most important -- or the largest -- in terms of spend optimization potential. If a marketer doesn't know what channels, tactics, and campaigns are performing better than others, it's hard to optimize spend in the right areas. Brands need to embrace a connected, multi-channel approach and measure efforts and performance across these channels consistently.
The tool doesn't even look at all search
In addition to being limited to search and overlooking the impact of other channels on a conversion, the tool doesn't even capture the entire universe of search -- it only collects data on Google search keywords and neglects keywords from others. What about the potential customers who are using Bing, Yahoo, and other search engines? Marketers using Google's attribution approach miss an entire subset of performance data to fuel decisions on marketing spend.
The tool is merely positional
The models that Google's attribution tool uses only measure results based on position, or the order in which the search terms are clicked, which makes it extremely unsophisticated compared to more advanced capabilities now available. By only measuring this dimension and omitting data on frequency, creative, etc., marketers are missing out on incorporating insights from other factors. In building an attribution model for our customers, position is only one factor that influences ad spend -- other factors can still be important.
The models are completely arbitrary
Finally, the tool requires marketers to select which model they would like to apply, rendering them completely subjective. Google's tool offers five different options: last click, first click, linear, time decay, and position-based. Marketers are then given the choice to compare three attribution models at once. But how would marketers know which model to choose? What logic would they use? In a more robust, algorithmic attribution approach, the correct and multi-faceted model can be built. This model is not chosen based on gut feeling or a marketer's idea, but by doing the heavy lifting of processing all the data -- for millions of converters and non-converters -- and discovering what works and what doesn't.
While Google's attribution solution is a step in the right direction, true attribution is a complex process that should be handled cross-channel -- tactic, publisher, site, etc. -- for the most reliable and actionable result. Attribution should not focus on strictly one channel, like search. And since Google has a stake in the ground on the publishing front (it sells keywords), it's not an uninterested party in the marketing pie. Given this and the factors listed above, marketers should instead consider an agnostic solution that measures and analyzes different tactics to uncover the whole marketing pie instead of just one slice.