Why Multi-Touch Weighted Campaign Influence Is Essential for Understanding Deal Attribution
- AUTHOR Luke Duncan
- May 21, 2014
- No Comments
Attributing revenue to campaigns has been a difficult challenge for data-driven marketers since the beginning. Throughout time, the tools and methods available to us have gotten more powerful and sophisticated, from single campaign attribution, to multi-touch, and most recently weighted influence.
The earliest methods would attempt to find a single campaign that could be linked to a deal, typically either the earliest campaign or the most recent campaign. Quickly, marketers realized that this one-dimensional method of attribution didn’t really reflect reality because nearly every deal involved multiple campaign touches. Giving attribution to just one campaign while the other campaign received none skewed the data in favor of specific campaigns that were more likely to be the “First” or “Last” campaign, but effective campaigns which didn’t were overlooked.
Faced with the issue of uncovering the rest of the attribution picture, multi-touch attribution was created. Essentially the idea is that instead of picking one campaign to credit, you give each campaign equal credit for the deal. This solves the issue of potentially missing a campaign, but introduces a larger problem. If you treat every campaign equally, how do you tell the difference between the email campaign touch that was sent to a random person at the company (and potentially never even read) and the webinar the decision maker attended (which was the only marketing touch he had time to consume)? In other words, if you make no distinction between a high-quality touch and low-quality touch, your data will skew towards high-frequency campaigns without much insight on their actual contribution to deals. That is where weighted campaign influence comes into play.
In most cases if you dissect a deal and look at all the campaigns and when they happened, all the people and the things you have identified about them including their role in the deal you probably would not choose to distribute attribution evenly. While you will never have perfect information, you likely already track enough to synthesize into a more accurate picture of campaign performance. If you want to incorporate the logic of first touch into a more holistic view of campaign performance you can weight it relatively higher than the rest of campaigns, at the same time you can ensure that campaign touches that key contacts have interacted with also get a relative bump. In addition to being a better way to judge general campaign performance, weighted influence also allows you to create models to answer specific questions like which campaigns are most effective at reaching influencers, are those campaigns different than those reaching decision makers, which campaigns are impacting deals early, and which are impacting deals just before it closes.
An attribution model, whether it is using a formula to identify a single campaign to give credit or crediting each campaign evenly, must always be evaluated based on the question you a trying to answer. An even-spread across all your campaigns may be a good place to start, but if you want your data to answer the question: Given everything I know, which campaigns have more effect on deals closing? You need to be able to translate what you know into usable aggregate data.