Multi-touch attribution: The fundamental to optimizing customer acquisition
For a long time, brands have dreamt of getting a bird’s-eye view of their customer journey, with the goal of gleaning insights into where their customers come from, how they interact with marketing channels and campaigns, when they convert and why. From those insights, brands can take action: doubling down on the tactics and channels that bring in the most valuable customers, and discarding those that burn cash and resources. As signs emerge that the economy is worsening and brands continue to tighten their belts, this bird’s-eye view of the shopper journey is more relevant and essential for success than ever.
Customer acquisition cost continues to skyrocket
Privacy changes in iOS and the lack of cookies don’t keep ads from reaching their target audiences. However, they do add difficulties to marketing attribution. Without Apple Identifier for Advertisers (IDFA) and cookies, ad platforms like Facebook can no longer tie a conversion to specific ads.
So how do brands solve the puzzle and discover the most effective places to put their marketing dollars to work? Before we dive in, let’s look at the following example to review what marketing attribution is.
Marketing attribution 101
When I first moved to the Bay Area, I wanted to buy an air purifier to deal with the possible air quality problem from wildfires. Knowing little about air purifiers, I started my research with YouTube videos, reading review articles and listening to what influencers had to say. Once I felt informed about what made a good air purifier, I went ahead and searched for the best options — comparing product specs from brands’ websites, Amazon, etc. Finally, I decided to go with Brand 2 as shown in the chart below.
Among all the touchpoints, Youtube Paid Video 1 and Reddit Review Article 2 drove my purchase from Brand 2. But how can Brand 2 know those two touchpoints moved the needle in my case? That’s where marketing attribution comes in.
In practice, marketing attribution relies on technology that:
- Collects shopper clickstream data on touchpoints
- Pieces disparate touchpoint data to construct a holistic shopper journey and
- Assigns credits for conversion to each touchpoint
Due to the technical challenges this end-to-end visibility often creates, most brands make marketing decisions based on first-touch or last-touch attribution alone. As a result, their marketing resource allocation is suboptimal at best. Luckily, machine learning provides a solution to multi-touch attribution.
Connecting the dots with machine learning
In the air purifier example, tens of thousands of shoppers interact with Brand 2 every day. At each touchpoint, the data system collects hundreds of shopper attributes and constructs shopper profiles. Each shopper profile serves as an identifier when IDFA, Android ID or cookies are not in place.
Later, machine learning algorithms leverage the shopper profiles, and automatically group shoppers who have gone through the same path into a cohort, as illustrated in Cohort 1, Cohort 2 and Cohort 3 below. Let’s say Cohort 1 brings in $1,000 of revenue. To understand Youtube Campaign B1’s performance, algorithms find all cohorts that have seen Youtube Campaign B1, and pair each cohort with a counterpart — where one touchpoint is different from the others — as illustrated in Cohort 1.1.1 vs Cohort 1.1.2 and Cohort 1.2.1 vs. Cohort 1.2.2.
In each pair, algorithms calculate the conversion delta and assign credits to Youtube Campaign B1. By iterating the process and summing Youtube Campaign B1’s conversion in each cohort, we get the campaign’s total conversion.
In the current economic climate, multi-touch attribution is more relevant than ever
Being disciplined and investing only on channels and campaigns that convert will help brands significantly as the economy is trending downwards and consumers have become more cautious about spending. With multi-touch attribution, brands can understand how each campaign influences consumer buying decisions, and whether the campaign is worth the marketing budget or not.
At the end of the day, brands invest marketing dollars to generate return. Channels and campaigns that are not converting need to be turned off. In the current economy, the faster a brand can identify what’s working vs not, the better.
In the following articles, I will continue to explore data-driven customer acquisition strategies. If you want to chat about them, feel free to contact me on Linkedin.
