How to Actually Improve Facebook Event Match Quality

In Summary, You need to understand if its actually possible to improve

After having set up your Server Side events, you will start seeing numbers or labels in the column “Event match quality”. These represent how effective the customer information parameters sent from your server may be at matching with Meta’s own database. Stakeholders will often question why these seem low and what steps are required to improve them.

There are numerous benefits to having a high Event Match Quality score, including seeing more conversions. But this is where the information can be a bit confusing, as the score does not reflect the quality of your Server to Server set up.

How is Event Match Quality Calculated?

To improve we first need to understand how it’s calculated, which is a closely guarded secret by Meta, but they do give us a general idea of what variables make up its score.

If we want to improve the score we need to focus on sending back the three parameters with the highest impact, including Hashed Email, Client IP, and Hashed Phone number.

However not all events can be improved. The data is required when the event takes place, but these parameters will only be available on some events. If we really want to improve the score we must first understand which ones can’t be improved.

Example: PageView event

A PageView event will fire on all pages. Unless your website requires the users to always be logged in, then an email or phone number will likely not be available to use. And you will be limited to only using the IP address (In the example above, this is the the first event with a score of 3.9).

Example: Lead event

A Lead event will fire when a user submits a form or books a call, we will likely be capturing an email, phone number, name and last name. Here there are a lot more opportunities to improve the score since the data is available to use (albeit we need to first transform it).

Case Study

We set out to improve the “Event match quality” score for a client on Lead events. The form contained an email and a phone number.

The hashed email was already being sent, so to get the biggest impact we added the hashed phone number.

What was the impact?

After this deployment we saw an increase from 6.2 to 7.3

Final Thoughts

It’s easy to associate the Event match quality score with the quality of the Server to Server implementation. Stakeholders must be assured that these scores only reflect the quality of the user data being received, and not all can be improved due to the nature of the event.