David is a Senior Analyst and has been at Wheelhouse since January 2016. After earning his BA in Politics & Government at the University of Puget Sound, he moved to the east coast to work at an analytics and scouting company for Major League Baseball. He has collaborated with clients across different verticals on search engine marketing, product listing ads, and display. He has worked extensively with e-commerce and B2B clients, and recently has focused on growing paid social advertising at Wheelhouse.
This is a guest post from David Creatura, Senior Analyst at Wheelhouse DMG.
One of the most important questions you can ask about your Facebook advertising efforts is, “How am I attributing our traffic and conversions?” Attribution is a commonly covered topic purely because of the wide range of opinions and perspectives one can have on the issue.
Customer journeys are often complex, and far from a one-to-one interaction between a single advertisement and conversion. Knowing how Facebook’s attribution works is critical to understanding your social advertising in a nuanced way to grow your program, versus relying on the default view.
How do we value a conversion for an individual who clicks on a Facebook ad one day, comes back to the site organically the next, opens an organic Instagram post a week later, and eventually converts via an email two days later?
In this scenario for our example company, the time between the customer initially clicking the ad and completing a purchase was 11 days. In the company’s analytics platform, the purchase would register in the default view as coming from the email campaign, as that was the last interaction the customer had prior to converting.
However, as the customer journey shows, there were many other touchpoints that helped move the customer closer to the conversion. Should we not also give credit to the organic Instagram post for keeping the customer engaged with the brand? Does the Facebook ad deserve full credit for being the initial interaction with the brand?
There have been plenty of publications focusing on the inexact science of attribution. No one model is a perfect fit for every company. Depending on your customer base, industry, brand, products, etc, there will be a different attribution model that will make the most sense for your advertising efforts and business planning.
Let’s use a fake e-commerce company to help explain the process. In honor of our resident furry friend in the Wheelhouse DMG office, we’ll make it an artisanal cheese manufacturer called Reggie’s Parmigiano Reggiano. We’ll use Reggie’s business to help us understand Facebook’s default attribution.
Reggie launches a Facebook ads campaign with the goal of driving a positive return on ad spend (ROAS) of 4:1. After a month of running his advertisements across different audiences he pulls in his “Website Purchases” and “Website Purchases Conversion Value” columns to see how his ads performed. (Ideally Reggie checks his campaigns routinely during the month, but he has other standard business priorities such as walks, begging for food, and angling for belly rubs.)
After pulling in the columns he sees that he drove $11.5k in revenue from just over 200 purchases and a total spend of $2.5k. After dividing his revenue by spend, he sees that he drove to a ROAS of 4.6 and exceeded his goal! This would seemingly signal the opportunity to spend more aggressively on his Facebook campaign, as he has room to hit his ROAS goal.
In this scenario, we would encourage Reggie to look a little bit deeper into the initial data that Facebook is showing him. Companies are often unaware that Facebook’s default attribution view is a 28 day post-click and also a 1 day view-through conversion. This means that Facebook will give 100% full revenue credit to an individual who clicks an ad so long as they complete a purchase within 28 days of clicking.
Facebook will also give 100% full revenue credit to a customer who just scrolls past an advertisement on their feed and then converts within 24 hours. This has significant implications as Facebook will show revenue for someone who does not click through an advertisement and will inflate the actual performance of your ad campaign.
This can be especially misleading if your campaign contains audience targeting of previous customers, email subscribers, or past site visitors. These are individuals who are already familiar or fans of your brand and have a higher likelihood of converting than a new prospect.
They very well may scroll past your ad while looking at their friend’s news for the day, then convert from an email that was sent as part of a broader campaign strategy. When presented with this scenario the Facebook ad is getting credit for performance when the email actually drove the revenue.
Now, this is not to say that view-through conversions can be disregarded, because there is legitimacy in accounting for view-through conversions. Thinking about attribution a bit differently than the default view, however, we recommend taking a percentage of that revenue to account for the impact an ad impression has in a customer journey as opposed to 100% credit.
This percentage will vary across advertisers, but if there is no standard that they apply for, it’s recommended that view-through conversions be roughly 20% of the total view-through revenue. We’ll touch on how to calculate the unique percentage for your advertising efforts in a future post.
To better help Reggie understand his artisanal cheese business success, we would segment our purchases by conversion type to change the way we view our campaign. To segment out by conversion type, click your column’s dropdown and select “customize.”
In the lower right of the available metrics to pull in, there’s a section titled “Attribution Window.” Beneath that, you’ll find Facebook presenting its default window of 28-day post-click and 1-day post-view. In our example below, we’ve changed that to shorten the post-click window for this specific client.
Upon editing your attribution window, the columns for “Website Purchases” and “Website Purchases Conversion Values” will expand to show the number of view-through conversions and revenue, as well as click conversions and revenue.
In our example with Reggie, we see that the ad campaign drove $6.8k in revenue using 7-day post-click attribution. If looking purely at click-driven revenue, this would result in an actual ROAS of 2.7.
That ROAS represents a 41% lower figure than what Facebook presents in its default window many advertisers are using to judge performance. Instead of increasing bids like Reggie initially would using the default view, Reggie should be significantly decreasing bids.
When we account for the view-through revenue, Reggie drove just over $3k in revenue using view-through conversions. Taking 20% of that $3k in revenue would show $600 in revenue is a fair representation of a view-through conversion. When added to our $6.8k in revenue that results in a complete revenue view of $7.4k, or a ROAS of roughly 3.0. At this point, Reggie still needs to pull back his bids!
To reiterate, no single attribution window or view is a one-size-fits-all for advertisers. But knowledge and further analysis about how Facebook presents your ad performance will tremendously alter the way that you’ll bid on your Facebook campaigns. For purely performance-focused campaigns and advertising efforts, this is a critical first step in ensuring your Facebook advertising campaigns are grounded in a reporting foundation that will allow you to scale your program.
