Putting Insights Into Action: Operationalizing Attribution
I recently spoke to an audience of search marketers on the topic of operationalizing the valuable results and insights that marketers and their agencies can discover through the attribution management process, and how frustrating it can be trying to translate these into repeatable, ongoing processes if you don't know where to start. Here are a few highlights from the advice I shared with them.
Don't Reinvent the Wheel
You undoubtedly have a set of critical success metrics or key performance indicators that you already use. These could include revenue, ROI, CPA, ROAS, LTV - any number of business performance measures that are important to you and other stakeholders within your organizations. You've not only operationalized the ability to track these metrics and report on them, but to make changes to the systems that impact their performance in order to optimize them. You have processes in place for changing bids, pausing keywords, channeling more funds to one publisher/network/search engine than another, etc.
So an important key to operationalizing the findings produced by attribution is to utilize those pre-existing systems and processes. But in order to do so, you have to adopt an attribution methodology that re-calculates all the success metrics with which you already work, to incorporate cross channel, cross campaign and cross attribute impact. When I say "attribute" I am referring to any "trait" associated with a touchpoint that your marketing efforts would produce with a prospect, such as search engine, keyword, ad size, placement, time of day, etc.
By having the critical success metrics you already use recalculated to incorporate attribution, you need only apply the processes you already use -- but utilizing these revised metrics -- in order to take more informed and effective steps toward optimization.
Identify and Act Upon Affinities
The attribution methodology you adopt should also identify affinities that exist between channels, campaigns and tactics. An example includes an affinity between a non-branded and branded keyword which, when run simultaneously produce greater ROI than when the branded keyword is run without the unbranded keyword. Another example would be the creative of certain online display ads driving searches, clicks and conversions on a particular keyword at a greater return than when the display ad creative is not active.
Once again, you already have processes in place to manage the changing, activation and deactivation of paid search ads and online display ads -- which can be used to greater effectiveness when affinities between certain attributes are identified.
Leveraging these affinities improves both your search and display results, with display serving as the more efficient bait that attracts potential customers and search serving as the more efficient hook that lands the conversion. This is why the idea that performing attribution results in devoting less budget to search is a fallacy - for if the affinities between search and display are used correctly, both search and display will end up getting more budget as a result of the improved performance of each.
Insist on Multidimensional Attribution
Just as no competent search marketer would base the optimization of a campaign solely on the fact that one search engine produces a better return than another - without also optimizing at the ad group level, the ad creative level, the keyword level, the day-part level, etc. - attribution also must not stop at the single-dimensional level. Despite the fact that you can get attribution metrics within publishers, or within ad creatives, or within keywords, in order to take full advantage of the power of attribution, as well as your existing processes for optimizing your efforts, you must insist on an attribution methodology that allows you to dig as deep and granular as possible to uncover the hidden jewels within your performance data.
For example, perhaps a given campaign has five different publishers, three different creatives and two different sizes. Stopping your analysis before you get to the most granular level that's available to you might result in your optimizing to a specific creative, when in fact there was one size of ad that produced far greater return than the other. Without having attribution performed in this manner you're flying relatively blind when it comes to optimizing on the attributes that really matter.
Give yourself some credit, your existing systems work -- and insisting on an attribution methodology that feeds them with more informed intelligence takes advantage of the best of both worlds.