Marketing attribution refers to the analysis and understanding of the touchpoints in a customer journey which lead to a conversion.
Customers have an increasing number of interactions with both online and offline channels before they “convert,” whether that be making a purchase, downloading an app, or simply signing up for a newsletter. For marketers, it is more important than ever to have a clear understanding of which steps in the journey are more or less valuable to the final conversion, and therefore more or less deserving of their budgets.
What is a marketing attribution model?
Marketing attribution models help businesses visualize how much value is in each step in a customer journey. Organizations who might not have considered attribution have likely defaulted into a simple measurement model (often ‘last-click’), but other models derived from data can be used to ensure more weighting is given to other touchpoints in the journey, according to the impact they had on the final conversion.
Marketing attribution models fall into single-touch and multi-touch categories. As we will see, some are more prevalent than others, and some are more superior at giving marketers clarity as to what steps in the customer journey are providing the best ROI. In doing so, they also highlight which touchpoints are deserving of more budget allocation.
As mentioned above, ‘last-click’ is often the default attribution model used. This is unsurprising given even a fairly simplified understanding of readily available web analytics software can see the last touchpoint a user has interacted with before making a conversion.
‘Last-click’ is appropriate in some cases (such as, when significant investment has been placed on that specific final channel in the journey). But as we know, customer journeys are increasingly complex and individuals are interacting with many different touchpoints across online and offline channels before converting. The ‘last-click’ attribution model, therefore, does not go far enough.
Fay Miller, Director of Marketing at Fospha, says “Last-click has been the industry standard since measuring customer journeys became possible. I usually see it used by businesses who have legacy reporting models, and who just don’t have a better way of measuring their marketing campaign performance. By “better,” I mean an approach that uses data to analyze all customer journeys and provide insights as to the role each touchpoint plays in the overall path to conversion.”
Another example of single-touch attribution modeling is ‘first-click’.
As the name implies, this model attributes weight to the first step in a customer’s journey – so might suit a campaign with an aim to create awareness – but ignores other potentially important interactions afterwards.
Business Development Manager at Fospha, Jamie Bolton, notes their relevance in some cases: “First-click is useful as a counterpoint to last-click. It is available by default in many platforms so it can provide another look into the data, particularly for organizations who have limited resources.”
The ‘linear’ attribution model is arguably the most straightforward multi-touch model.
Marketers using this model are acknowledging that their customers are interacting with multiple touchpoints on their journey. The linear approach credits all channels equally, so while it is certainly an improvement on single-touch models, it fails to account for the fact that different touchpoints may have had more or less of an impact on the customers decision to convert than others.
“It is a great logical progression from single-touch models,” Bolton says. “Linear attribution opens up the full customer journey and assigns credit evenly across it, meaning touchpoints that play a discovery, nurturing, or converting goal are all rewarded.”
‘Time decay’ is another multi-touch attribution model. It credits all touchpoints on a user journey, with an increased weighting the closer the touch is to conversion.
The value of the ‘time decay’ model, and its flaws, are quite easy to quantify. Of course, there will be many customer journeys where channels visited closer to conversion will have been more impactful than those near the beginning. But as user journeys get longer and more complex this is increasingly not the case.
‘Time decay’ still involves a fair amount of guesswork and an assumption that channels visited near the conversion deserve more credit than early ones.
Bolton highlights that ‘time decay’ has value for marketers who are focusing on quick conversions. “For businesses that want to focus on touchpoints that play a converting role for their customers, this model considers and values all channels, but prioritizes those that directly prompt conversion.”
“U-shaped” – or “position-based” attribution is another useful multi-touch model which is a clear improvement on the single-touch methods explored above.
This model does credit all touchpoints, but gives more weight (40%) to each the first and last channels. The remaining 20% of credit is spread across channels in the middle of the journey.
Again, the “u-shaped” model has relevance, but it can fail to accurately credit any touchpoints in the middle of the journey which may have had a bigger part in the conversion than the marketer might expect.
As Bolton notes, “U-shaped is a strong candidate for longer customer journeys. The model works on the principle that the discovery step and converting step contribute the most to a sale, but doesn’t entirely discount the role played by nurturing steps in between. In a customer journey with 10+ steps, a linear attribution model would dilute the value of the first and last step until they are almost not recognized. Businesses that want to avoid this can use a position-based model.”
‘Data-driven’ attribution is another multi-touch model. But unlike those we have discussed so far, it uses data across touchpoints to eliminate any guesswork and to attribute credit to channels according to their performance, rather than by what position they are in.
“A data-driven approach is often better than other measurement models because it uses journey data to analyze customer behavior and the route to purchase,” says Fospha’s Fay Miller. “It provides insight as to which touchpoints are outperforming the rest. What’s more, it adapts to what the historical journey data insights show and will adjust based on the performance of particular channels.”
As the name suggests, “data-driven” attribution calls on having comprehensive data and a full view of the customer journey, along with the means to analyze and act upon that data effectively. In 2016, Forrester reported that as much as 73% of all data within an enterprise goes unused. Businesses need a Customer Data Platform in place to consolidate, organize, and as Miller puts it, “automatically assign value to channels and tailor the marketing strategy in real-time.”
“Where budgets permit,” she continues, “data-driven models should be the first choice for any marketing team. While more costly initially, this spend could easily be offset by higher ROI resulting from the campaign.”
Example of data-driven attribution
A basic campaign for your new product might involve organic search, display advertising, email marketing, and social media posts. If your display ads are the first touchpoint, and organic search is the channel used before conversion, ‘data-driven’ attribution would be better placed to allocate more or less credit to the emails and social media posts that might appear in between.
It’s possible that, for example, the emails your customers are seeing are prompting more eventual conversions after the final organic search step – and so a ‘data-driven’ attribution model would give more weight to this channel over the social media posts.
More businesses switching to data-driven models
We can see that with access to data from all channels across the above example journey, this model is going to offer more accurate attribution than other multi-touch and single-touch models.
By comparison, a “first-click” or “last-click” model would ignore three out of the four touchpoints. A “linear” model would still over-credit some steps and under-credit others. “Time decay” would likely under-credit the value of the awareness-generating display ads. And a “U-shaped” model wouldn’t offer the control over shifting weight between the two middle, email and social post, steps.
It is understandable, then, that senior marketing leaders are increasingly devoting more initial budget to “data-driven” attribution. The added insight from comprehensive, and often readily available, data is enough to make savings from better ROI in channels you can be sure are deserving of more budget allocation.
With comprehensive data and the tools to unlock it, multi-channel marketing can be more accurate and agile, perfectly suited to today’s world of increasingly complex customer journeys.