Main content

Marketing attribution done right - some best practice tips

Barb Mosher Zinck Profile picture for user barb.mosher May 3, 2022
Marketing attribution is needed to identify how well - or otherwise - a marketing campaign is doing. Here are some ideas on how to achieve that!


There is no easy way to approach marketing attribution. However, there's also no need to get caught up in first-touch, last-touch, multi-touch attribution conversations to learn how to prove the value of our marketing efforts.

This idea became very clear when I listened to the MarketingProfs B2B Spring Forum session by Moni Oleyede - 'Attribution Done Right'. Oleyede is the Director of Marketing Infrastructure at Fidelis Cybersecurity, and she has extensive experience working on all types of digital marketing programs.

Oleyede started her  session by saying that attribution is hard. But it's not the model that's hard; it's how we perceive it. She argued that we collect plenty of data, but we don't trust it. We also create complex data structures that miss the mark.

The mistakes we make with attribution

There are a few mistakes marketers make when building their attribution models, Oleyede said:

  • We try to put everything in one model, and it's too much to understand and figure out
  • We don't create clearly defined goals. "ROI is not a goal."
  • We think there is a perfect process, but there isn't
  • We depend on a linear model (like a lead funnel) with predefined timeframes. Oleyede said the predefined timeframe is the biggest problem (e.g., one month, one quarter). If you have to use a timeframe, she said, compare it to another, so the timeframe is in context. So, for example, look at this spring versus last spring or this holiday season versus the prior one.

Oleyede said that the number one thing that is ruining attribution models is trying to track revenue. Revenue is a business goal she said, not a marketing goal.

Setting a goal like what is working or generating the most ROI isn't going to work. These may be valid business questions, Oleyede said, but they are too big to answer with a single attribution model. They also require data outside marketing to answer.

Your attribution model needs to show how your marketing efforts influence a person to take a specified action. A single goal is what you need to focus your model on answering.

Building the attribution model

So how do you create an attribution model that works and shows you how a marketing campaign or program is doing?

If you want to get attribution right, you need to focus on answering questions that marketing can control, and it should be based on customer metrics, not business metrics, Oleyede said. What matters most to your customers? What motivates them, drives them to a free trial, and what turns them off in a customer experience? Once you understand your customer, you can figure out the questions you want to answer and the KPI to track with your attribution model.

You also need to add context to your data. Without context, analytics don't mean much. You apply context by mapping your model to the buyer's journey and looking at the results of activities in terms of how well people are moving through that journey. With the buyer's journey, you map out the stages a person moves through to reach a goal, including what they are doing at each stage, what content they are looking for, what's the next step you want them to take to move them to the next stage, and what are the success metrics for that stage.

Oleyede provided the example of a first-time home buyer looking for an agent to help them buy a house. In this example, the goal is to get the home buyer to fill out a Request for an Agent Form. She mapped the buyer's journey from Interest (Awareness) to Consideration, Validation, and Decision, including what content is offered at each stage, the desired next step to move the person to the next stage, and the success metrics measured at each stage.

When you build your attribution model following this approach, it's easier to understand what is working: what content is converting, what new content you may need, where are people getting stuck or dropping off in the journey, where you need to focus your efforts to get the conversion, and so on.

The point that Oleyede made is clear - if you focus on customer metrics and build your attribution model around those metrics, the ROI takes care of itself.

My take

What I like about this approach to attribution is its simplicity. Not that it's simple; there is still a lot of work to create the attribution model, but the process is straightforward. Understand your customers and what drives them, map the buyer's journey to the goal you want to achieve, identify the content and steps you will take at each stage to keep the buyer moving forward, measure, and improve.

Oleyede pointed out that not all marketing KPIs are hard numbers you can track. Brand awareness and loyalty are two examples where there is no end "hard" KPI to measure, but instead, a number of metrics together can give you a clear understanding of how those KPIs are doing.

The other thing we have to consider with this model is that buyers don't always buy in a linear journey; they tend to jump in and out of stages or look for different content in a stage we wouldn't expect. So we have to plan for this by paying attention to how the journey works for most people but preparing for outliers that can influence how we improve our attribution model.

Nothing is ever set in stone in marketing. We need to continually review and adapt to keep up with evolving customer expectations. I appreciate that Oleyede has provided an approach to attribution that can help us do that.

A grey colored placeholder image