Attribution may be the single most important element of advertising. It not only helps advertisers see what works and what doesn’t by assigning conversion credit to ads, but it impacts optimization and targeting. Get this wrong, and your mistakes spill down through every step of your advertising.
Last week, we covered the most common optimization mistakes that advertisers make. This week, let’s focus on attribution mistakes.
Maybe you are making some of these mistakes. It’s not too late. Make the necessary corrections.
Let’s get to the most common attribution mistakes…
What is Attribution?
First, let’s be clear about what we’re talking about.
Attribution is the ability to give credit to an ad for a conversion. While a simple concept, there are several layers to proper attribution that can impact advertising performance.
When something goes wrong with attribution, it’s due to a failure in one of these areas…
1. Setup. You’ve done everything on the back end to make sure that results are reported accurately and completely.
2. Application. You know how to apply your knowledge of attribution to different optimization strategies.
3. Interpretation. You are able to make meaning of your results.
4. Understanding. You understand how attribution works, it’s strengths, and weaknesses and how that impacts your approach.
Now let’s break down the most common mistakes.
1. Improper or Incomplete Setup of Pixel or Conversions API
Conversion attribution is impossible without first setting up a way for Meta to know how people are engaging with your business away from the Meta family of apps.
This starts with the Meta pixel. It needs to be on every page of your websites. When possible, it needs to be on other websites you don’t own where you sell products.
At one time, setting up the pixel (and events, which we’ll get to in a moment) was enough. But mostly due to privacy laws and weaker tracking, passing first-party data is critical to complete attribution. If you haven’t set up the Conversions API, you are sending incomplete data to Meta.
That could come in two forms:
1. Web API. This is the most common form of the Conversions API. By sending conversion information from the web API in addition to the pixel, you can help fill in blanks where the pixel can fail. There are multiple methods to accomplish this, but I use Stape to set up the API Gateway.
2. Offline or CRM API. If all business is done on your website, the web API may be sufficient. Otherwise, offline leads and purchases need to be passed to Meta via an offline or CRM API. This allows the possibility that you can see when your ads lead to conversions that happen away from your website. Meta can also optimize for these conversions.
2. Failure of Standard Events, Custom Events, and Custom Conversions
Of course, setting up the pixel and Conversions API is half the battle. Make sure you do that. But it’s the events themselves that define when someone performed an important action.
Events help define whether someone performed a purchase, registration, search, or other important action on your website. Custom conversions help provide granularity to your reporting, like the specific product that was purchased.
Failure in this area comes down to three primary things…
1. Misunderstanding their roles and unique purpose. Do you know the differences between standard events, custom events, and custom conversions? Most advertisers don’t, confusing custom events for custom conversions. Advertisers will attempt to use one in place of another. The reality is that you need to use all three.
2. Improper or incomplete setup. Set up standard events for all important actions when possible. Set up custom events for those unique actions that aren’t predefined. Pass the necessary details of these actions via parameters. Create custom conversions to add granularity to your reporting.
3. Over or under reporting. When results are clearly wrong, advertisers are often quick to blame Meta. But start with yourself. The pixel, Conversions API, and events all need to be set up properly to fire on the right page and at the right time. Do this incorrectly, and you may send too many or too few events, which will impact your reporting.
3. Inability to Understand Meaning of Conversion Results
One of the most fundamental failures is misunderstanding your results, how they are calculated, and the context behind different types of attribution.
By default, conversions are counted when someone clicks your ad and converts within seven days or views your ad and converts within a day (without clicking). Far too many advertisers have no idea this is the case. They assume that all reported conversions in the Results column are due to someone clicking their ad and immediately converting.
That conversion may not be immediate. It may happen later that day. It may happen seven days later. Or your customer may not have clicked at all, but they were shown your ad.
Attribution mistakes often come down to misunderstanding that either all conversions are equal or that all conversions of a type (1-day click or 1-day view) are always good or always bad.
Context matters.
If you are an experienced advertiser who appreciates the nuance of the various types of attribution, you regularly use the Compare Attribution Settings feature to see how your results break down. You’ll even add a column for 28-day click, which is otherwise buried.
How many of your conversions are view-through? Depending on what you’re promoting, a high percentage is a red flag. You may want to discount them. Or simply acknowledge that they aren’t as meaningful as the the click-through conversions.
Of course, if you’re selling a product and a high percentage of those view-through conversions are engaged-view (and your ad uses video), you may have more confidence in those numbers.
There’s also the matter of visitors performing a conversion event multiple times, which can lead to the perception of inflated results. This can be addressed with First Conversion reporting.
4. Expecting Google Analytics and Ads Manager to Report the Same
One of the advertiser’s biggest annoyances is a client who insists that Ads Manager reporting is wrong because it doesn’t match up with Google Analytics.
How do you respond?
Meta and GA4 will measure your conversions differently. And frankly, Google has less data than Meta does.
Only Meta has the knowledge that someone saw your ad without clicking prior to converting. And Meta may be better equipped to attribute a conversion to an ad when a customer switches devices or comes back days later to complete a purchase.
It doesn’t matter that you use UTM parameters. This still doesn’t solve for view-through conversions. And it’s unlikely to be enough to help GA4 properly attribute a conversion from Facebook if it happens days after the initial click.
It’s important to use both. Use GA4 with UTM parameters as a second source of information. This can also help you spot problems if you are unable to explain the disparity.
But one isn’t “right.” Neither is perfect. Embrace this.
5. Always Leaving the Attribution Setting at the Default
A big mistake is misunderstanding how the attribution setting applies both to default reporting and optimization for ad delivery.
Once again, the default attribution setting is 7-day click and 1-day view. Not only does that mean that conversions will be reported that happen within that window, but Meta will optimize to show ads to people who are likely to convert within that window as well.
This is important. If you’re optimizing for purchase, a 7-day click and 1-day view attribution setting makes sense. But it may not for any other type of conversion.
You can make the case that a view-through conversion is relevant for purchases. Someone saw your ad. They were interested. But it is a big commitment. They need to discuss with their spouse, business partner, or higher ups. They either go directly to your website or Google your product later that day and convert.
But this explanation for view-through conversions falls flat when discussing the typical lead. If something is free and easy to acquire, it makes very little sense that someone wouldn’t simply act on that ad when they see it.
This also applies to when optimizing for custom events based on engagement actions. These events can happen repeatedly. The result is that Meta can inflate your results by simply displaying ads to people who visit your website regularly. Even if they don’t click.
There is a solution. Edit the attribution setting in these cases to be 1-day click only. Since view-throughs won’t be counted as conversions by default, the algorithm won’t optimize for that type of conversion.
You can still see 1-day view conversions, though. They just aren’t included in the default reporting. Use the compare attribution settings feature to see them. Expect that you’ll get fewer of them when they aren’t included in the ad set attribution setting.
Watch Video
I recorded a video about this, too. Watch it below…
Your Turn
What areas of attribution do you struggle with?
Let me know in the comments below!