Are you seeing positive numbers from your paid advertising efforts, but no correlation between that advertising and topline revenue? You are probably experiencing over-attribution.

The Facebook Ads and Google Ads platforms generally work in a vacuum and are apt to over attributing revenue dollars. This is because these platforms are the main revenue streams for these businesses, and as such, they want you to spend as much as possible. With that, attributed dollar amounts can easily be overblown, and lead to relatively flat business growth.

How does this happen and why is over-attribution so rampant in advertising platforms? Mostly because these platforms do not attempt to distinguish new and returning customers. 

The rise of the Facebook and Google algorithm has lead to advertising models where you try to gather as many conversions as possible, generally at the lowest cost. The idea here is that you are attempting to provide as much data to the algorithm as possible so that the algorithm can go out and find users who are apt to purchase. However, the user that is most apt to purchase, is generally a user that knows the brand and was going to purchase anyway. Feeding the algorithm with more and more data from users who were already going to buy places you in a cycle that turns even the most robust campaigns into large remarketing baskets.

HOW TO IDENTIFY OVER-ATTRIBUTION

The easiest way to determine over-attribution is to compare Facebook and Google Ad spend to topline business revenue. For example:

SHOPIFY REVENUE

FACEBOOK AD REVENUE:

GOOGLE AD REVENUE:

In this case, total advertising revenue is $1,985,893.23, while total website revenue is $2,030,394.43. This would mean that advertising made up 97.8% of all revenue. 

There is little to no chance that advertising made up that large of a portion of revenue. At most, we like to see paid advertising make up 60% of all revenue, and that is only during times of heavy growth. 

HOW TO COMBAT OVER ATTRIBUTION

The key to combatting over-attribution is to separate out returning customers and users who are familiar with the brand, with users who are truly new to the brand. This is easier said than done because it requires the user to recognize deficiencies in both the Facebook Ads and Google Ads advertising platform. 

Let’s take a quick look at both Facebook and Google Ads, and how you might put this model into practice.

FACEBOOK ADS

Here at Slicedbread, we make heavy use of both exclusions both from “Customer Lists” and “Website Purchasers” from pixel data. Our recommendation is to look at Facebook campaigns at any level of the funnel separately and to measure results independently of each other. This is totally different from the usual tactic of looking at the platform holistically. We recommend splitting and viewing as such:

• Top Funnel - Users who have never experienced the brand
• Exclude all past purchasers all-time
• **Exclude all website traffic all-time**
• Mid-Funnel - Users who have visited the site but never purchased
• Exclude purchasers who have converted in the last 30 days
• Low-Funnel - Users who have hit the site over the last 30 days or less
• Target only those users who have visited in the last 30 days and those with buying intent. 

The idea here is to isolate the top of the funnel to only those that are new to the brand. The hope here is to allow for the budget to be spent only in that segment to force brand expansion (profitably). The hope here is that you’ll be better able to control your budget and cost rather than overspending on users who are going to purchase anyway. 

GOOGLE ADS

For Google Ads, we also make use of customer lists in the form of Customer Match, as well as exclusions from the Google Analytics audiences (namely past purchaser lists).

Note here that Google Analytics audiences collect data from the 30 days from when they are created, so if you do not have custom audiences based on engagement set up, you will want to do so immediately.

The key to proper attribution in the Google Ads platform is to know Google Ads limitations. Search campaigns can easily house exclusion audiences, but Google Smart Shopping campaigns cannot. If you are not part of the Google Beta Programs (Shameless plug here: we are included), there is no way to differentiate between prospecting and remarketing audiences. As such, you are always going to be at the mercy of Google in this case, and Smart Shopping will usually have an overstated ROAS because of it. 

EFFECTS OF COMBATTING OVERATTRIBUTION

If you have completed the segmenting of audiences correctly, you are most likely going to see your ROAS fall. The biggest thing to remember from this exercise is that a reduced ROAS does not mean that you are seeing decreased results. What you have done is changed the way you advertise, and generating new business is generally more expensive. 

Once you have found your new ROAS baseline, you can more easily optimize for new customers. This will allow you to have more flexibility to add more to budgets and see the effects in total revenue. Inversely, it’ll also allow you to feel where your topline limits are, and when advertising effectiveness starts to wane. 

Holistically, you should be able to reduce remarketing budgets and you should also see improved ROAS in remarketing campaigns. Additionally, the increase in top-of-funnel Facebook marketing will cause Google Shopping performance to improve as well.

Hopefully, you now have a better understanding of how to improve your overall results just by restructuring.

Those are the basics for combatting over-attribution. Happy Advertising!