Attribution – the age old marketing conundrum. How do we measure customer engagement, and attribute dollars back to the behavior of the target audience? On one hand, it’s a pragmatic approach that makes marketing more empirically measurable. Behavior patterns can be measured and tied to dollars and cents, and it bridges the gap between the creative and the scientific, making Marketing accountable to Sales, and aligning the marketing team more closely to the sales force. On the other hand, however, it can also be viewed as forcing math.
Marketing’s first task is producing quality leads with a high probability of converting into real business for Sales. It is responsible for not just demand generation however, but also pipeline – businesses naturally want to understand just how much Marketing is contributing to revenue generation. By using Lead Scoring to identify pipeline progress, Marketing arguably loses some level of control over lead quality due to relying on algorithms to figure this out, but when attribution enters the picture, control is theoretically granted back to the marketer. Marketing now has visibility into which channels and tactics are most likely to drive positive impact to pipeline, and how this is applied for various segments and stages of the lead & customer lifecycle. But here is where the complexity lies. We are effectively attempting to measure the great variable that is human behavior and at the same time identify the types of marketing activities which can influence purchase behavior in the most beneficial manner.
Attribution models can often be misleading and not truly indicate the performance that Marketing is achieving. Fixed models focus on only one touch, and can’t possibly reflect a multi-touch pipeline. In general, mixed models assume a first touch, last touch, or some combination of touches leading up to opportunities. The problem with this approach is that the value/weighting of specific touches can be defined arbitrarily. That is, a business can choose to take more value from the first touch (or the last, or any touch!) and apply more credit to that specific touch. They are thus assuming that a specific touch is more valuable than another, with minimal to no data to support those claims. Structuring goals, KPI’s and even marketing plans around this is a dangerous proposition.
To avoid this situation, there are 2 options:
- Build a predictive model and supply it with sufficient data to train the model and apply it to the business. This is the only way to truly look at all behaviors and impacts holistically, but it necessitates a multitude of definitions, controls and data unification. This approach requires data science, broad scale integrations and, ideally, a central data store for a full customer view in order to build a data-driven methodology, a luxury that not all organisations can afford.
- Find an alternative solution – yes! This is for businesses who recognize the pitfalls of attribution, and have lived through the misleading metrics an inappropriately configured attribution methodology yields. Focus can be shifted instead towards constructive, immediately measurable marketing techniques, such as CRO, Marketing-Sales alignment, and ABM. Done well, any of these three will deliver positive impacts to the organization. They are all measurable without the need for major assumptions, and they yield immediately measurable performance.
In truth, developing a model, whether fixed or mixed, and then attempting to understand how much credit a whitepaper download gets for driving ROMI is fudging numbers a little. So is there a middle ground? Yes. There are a multitude of ways in which companies can strive for a full, data-driven view into attribution via a number of intermediary steps. For starters, an organization will need to define just what they want to attribute. Of course, Marketing will always want to credit campaigns for any revenue that is generated, but aiming for attribution with a laser focus on dollars takes away from other core value adds that Marketing brings to the table: building awareness, growing the database, driving advocacy, building a competitive edge through content and messaging—and many others! Thus, it is imperative that when Marketing attribution is used, it must be understood across all stages of the funnel.
At the very least, an organization should define and measure the following:
- What are the campaigns responsible for driving visitors to the site? Which campaign types are most effective at doing so? These are the awareness campaigns which drive interest and inbound visits.
- What are the campaigns responsible for converting visitors into known and marketable prospects? It is these campaigns that are responsible for creating marketing contacts and growing the reachable contact database.
- Which campaigns have the greatest propensity to convert known prospects into MQLs? Are there specific campaign types that are responsible for driving lead score up due to driving more desired behavior and data sharing from the prospect?
- Which campaigns show the greatest success rate in driving bookings? Which of these campaigns impact funnel velocity in a desirable manner?
…And there are many more than that! Imagine the above applied to a variety of geographies, customer segments, net new customers vs existing customers etc. It gets complicated fast, and requires careful scrutiny and planning in order to roll out appropriately, but is easier to implement than a full-blown Attribution model.