Why does my Google Analytics numbers not match Facebook Ads Manager for a Facebook Ads Campaign?
The combination of running Facebook Ads with Google Analytics as a measurement solution is one of the most common approaches to paid advertising. But with it comes one of the most commonly asked questions in ad tech: why does my conversion numbers not match between Google Analytics & Facebook Ads?
Truth be told, there will never be a possibility where these numbers will match but it does not seem to be very common knowledge in the digital ecosystem.
So let’s look at some of the major reasons why there are differences. Note I call these differences & not discrepancies because these are completely expected outcomes based on the two technologies.
Click vs Impression Tracking
The most straight forward explanation for the differences seen comes to how the two platforms measure attributed success for a conversion event.
Google Analytics (and a large proportion of GA360) relies on a click based approach to measure & ultimately track success. This is done in two ways:
UTM Parameter tracking, applied to the end of a click-through URL on a Facebook Ad e.g example.com?utm_source=facebook&utm_campaign=flashsale
Auto Tagging, which can come from the application of Campaign Manager (CM) click tracking on a Facebook Ad e.g. ad.doubleclick.net___?example.com?dclid=1234abcd
The latter does require GA360 for maximum effect and the application of Enhanced Attribution within the CM adserver. But allows users to get data in GA without applying UTM tracking, similar to how Paid Search works on a Google Click ID (GCLID) via Google Ads.
On the other hand, Facebook Ads Manager reports on both click and impression conversions by default. This additional impression level measurement can often be between 70-90% of all conversions on Facebook, which is something Google Analytics would never have a view on. And this leads onto the next core point.
Attribution Windows
By default, the Facebook Attribution model includes 1 Day Post Impression in it, alongside 28 Days Post Click. You can relatively easily override this on the Ad Account level in Facebook, as well as break out reporting into different windows.
Google Analytics on the other hand defaults to a 30 Day Attribution window as a default. Considering the maximum Attribution Window in Facebook Ads Manager is 28 Days, there is a little gap here which can lead to the differences seen.
You may be thinking, what if I just amend my Facebook Ads reporting to just report Post Click conversions only? Surely that would make it closer to Google Analytics numbers?
The answer is no, and there are a couple of reasons why.
Attribution Counting
As already explained, Google Analytics uses the click based tracking of a UTM or Autotagging to do its attribution. When an user lands on site from a Google Analytics tracked landing, the user is defined by the Client ID (or User ID if set up within Google Analytics) and then the Google Analytics pixel (gtag.js) goes to work to tie this Campaign data to any proceeding event/goal. The conversions are attributed to the day of conversion.
For example, if I clicked on a Facebook Ad with UTM tracking on 1st January but performed a conversion event on the 5th January, the conversion would show in Google Analytics reporting on the 5th January.
Conversely, Facebook Ads attribution works to attribute conversions to the day of impression/click. In the same example above, the Facebook Ads Manager reporting would show the conversion on the 1st January.
This is a big point to remember and it makes sense if you think about it. For example if you stopped running a paid Facebook Ads campaign, it is normally to see data still come into Google Analytics several days after from a conversion perspective, based on the methodology described above. Whereas in Facebook Ads Manager, you will never see data after the last day of spend. Which means reporting continuously changes backwards and it is important to re-pull this for the true results of your Facebook activity within your Attribution Window.
Core Attribution Principles
Another big reason for a difference is perfectly explained by defining the two technologies. Google Analytics is a web analytics tool by default (though is beginning to move into the app space also in place of Google’s SDK Firebase with it’s App+Web Properties), made to measure all interactions on site. Facebook Ads Manager is a buying platform for ads within the Facebook ecosystem, tied to a specific Ad Account & Data Source when running advertising.
Google Analytics can see all channels that you tag up with UTM tracking, Auto Tagging or nothing at all. Both Direct & Organic traffic can be seen & used in Attribution models within Google Analytics. Google does this by crawling the referrer URL from every page entry and combining it with any query string (UTM tracking) to determine the channel. You can change this by configuring the Default Channel Grouping setting in Google Analytics.
On Facebook Ads Manager side, the only activity it sees is itself. The Facebook Pixel that largely powers any conversion measurement on websites can see more than just Facebook but within Ads Manager, the only consideration are the Ad Accounts & Campaigns you are running with.
Therefore this creates a big dynamic, that explains further another difference in the numbers.
Let’s take the following example:
January 1st I serve a Facebook Ad from Campaign A but don’t click on it or convert
January 2nd I serve a Facebook Ad from Campaign B & click on it but don’t convert
January 3rd I click on a Paid Search Ad via Google but don’t convert
January 4th I see a Display Ad via Programmatic but don’t click on it or convert
January 5th finally I go to the website & convert
Starting with Google Analytics, it can see steps 2, 3 & 5. With GA360 and using CM as your adserver, it can also see step 4. Remember Google Analytics by default is click based hence not seeing Step 1 or Step 4. And without GA360, you cannot change the attribution model From an attribution perspective, 100% of the attribution would go to the Direct (Step 5) in a Last Click model or Paid Search (Step 3) in a Last Click Non-Direct Model. None of the Facebook touchpoints are valued here.
However in Facebook Ads Manager, it has no idea that Steps 3 - 5 have occurred, with the only exception that within Step 5, the conversion being triggered should also trigger the Facebook Pixel and it will attribute it to the Last Click event of Step 2 within Facebook’s world. When in fact, it was not truly the last channel to influence the path in the Last Click Attribution model.
So what does mean in conclusion?
Google Analytics will always factor other paid channels (that are tracked) and organic/direct traffic into consideration to determine who gets the credit for a conversion. GA360 fuels this further with the addition of other attribution models & the use of CM data if utilised. Facebook Ads Manager only sees itself and theoretically overvalues itself as a result in attribution.
As a small aside, it should not be forgotten that Facebook do have another solution for measurement that factors in other channels in a similar methodology to Google Analytics. This is called Facebook Attribution and you can read further on this via my blog post here.
Cross Device
Another big piece of the answer comes down to the idea of Cross Device, the view of the same user across both web & app. Facebook arguably have the strongest cross device solution within the entire digital ecosystem, considering it is a closed, logged in environment. Google on the other side may have more coverage across their properties but by default is cookie based. And in this specific situation, Google Analytics is very much a technology powered by cookies in its current, most utilised form.
So when we think of running a Facebook Ads campaign, it is important to note that Facebook Ads will use the power of Facebook’s Device Graph to attribute conversions to people. This is especially true when we talk about the post impression conversions mentioned earlier.
On the other side, considering we know that Google Analytics relies on a click from a Facebook Ad to even consider conversion counting, combining this with it’s limited approach to Cross Device means that a lot of conversions are often lost within Google Analytics. In principle, Facebook is more accurately able to attribute conversions to a single person and therefore has higher numbers vs Google Analytics.
Saying this, as previously touched upon, Google Analytics are now making a bigger push to close the gap on this specific area of Cross Device by the introduction of their Web+App Properties. And this is something that is going to increasingly grow into 2021.
The definition of a Click
To now go full circle, we will further examine what a click means. As we have seen so far, a click is super important to Google Analytics in comparison to Facebook Ads Manager from an attribution perspective. But it is also true that a click will not always equate to a a land on a website.
Both Google Analytics and Facebook Ads Manager tries to define this in different ways. Google Analytics uses the idea of a Session to identify if an user has stuck around and not bounced. Facebook Ads also has a metric called Landing Page View which uses the Page View event in the Facebook Pixel to tie to a click on a 1 Day Post Click Attribution window.
If your click to land ratio is not great, this can be another reason to consider why the numbers between the two platforms may not match. Whilst it is impossible to see a 100% click to land ratio, some reasons that could explain the drop off could include:
Latency of connection/signal
Too many redirects
Tracking breaking, especially if UTM tracking falls off (can be website or ad operations at fault here)
Slow website
Delay on pixel fires
It is always super important to click test your ads and make sure that the Google Analytics tracking from a Facebook Ad is coming through as expected. The golden rule of using a question mark “?” to append a landing page to any analytics tracking is important here, with the exception of where a question mark already exists so using a “&” is the solution.
But actually the biggest consideration in this specific section is how Facebook defines a click from an attribution perspective. And this may be the most shocking point in this whole article.
Facebook Ads Manager actually attributes based on All Clicks. Which seems plausible. But those social experts out there will know that the definition of All Clicks on Facebook is much wider than just clicks to a website. These are normally defined as Link Clicks on Facebook. All Clicks includes pretty much any other interaction, including things like clicking a share button or opening an Instant Experience Ad. And Facebook uses these clicks from an attribution perspective. Whereas with Google Analytics, a “link click” is a requirement.
This is a very unknown point in general within the ecosystem and has wider implication on measurement / attribution. Facebook do actually say this in their support documentation but it is not easy to find. You can find it here on the second paragraph in case you don’t believe me! But again, yet another reason to explain the gap in conversions between the two platforms.
Data Legislation
I couldn’t go an article without talking about the exciting world of data legislation. Whilst this is only relevant to certain countries that abide to GDPR & CCPA, this is increasingly a very important variable in digital advertising and again plays a role in this very situation.
By default, users should now only be tracked if they have given consent to the end vendor based on the logic of GDPR & CCPA. This is important as ever when it comes to adtech. But as Google Analytics and Facebook Ads are different types of tools, it is important to note what it means here.
Starting with Facebook Ads Manager, as this is a buying platform, the consent is largely opted in from the moment you have an account with Facebook/Instagram etc. They use this to serve ads with no real pressure. The only complication is more advanced targeting, specifically from a 3rd party e.g. the use of CRM Emails as a Customer List Custom Audience in Facebook. Facebook put all of the consent collection on the advertiser/agency to do this type of targeting. When it comes to measurement & conversion tracking however, the Facebook Pixel is more important here.
Which brings us onto Google Analytics. This is an Analytics tool by default and would fire on any website where the tag has been placed. Theoretically this now requires consent to fire, which is where the cookie notices / consent management platforms (CMP) come in. These generally can be configured to fire different on certain types of tracking such as:
Essential - tags that would break the site if not loaded
Analytics - for the use of web analytics
Behavioural - for some form of segmentation but not specifically meaning advertising
Advertising - tags that help inform ads, which is where the Facebook Pixel would fall in
There is an interesting debate of whether Google Analytics is defined as an Analytics solution or whether it is more a Behavioural / Advertising solution. Theoretically if you configure Google Analytics to its maximum and especially with GA360, it is more the latter.
So what does this mean for the conversion difference question? Well if an user opts out of one or both of Google Analytics / Facebook Pixel but still converts, there will be holes in the reporting. This is again really important to consider, that with opt out rates always going to be a tangible number, that this has a direct impact on attribution of digital advertising.
Conclusion
Hopefully this deepdive into the two systems of Google Analytics & Facebook Ads Manager has helped explain why it is impossible to ever see conversion numbers match between the two.
Some of these are very common knowledge and some are less so.
The ultimate question then becomes which is the true source of truth for Facebook Ads. Is it Google Analytics or is it Facebook Ads Manager?
In my opinion, both have pros & cons but actually neither one does the job. It is something that Facebook Attribution as a solution is looking to help paint a better picture of the role of Facebook Ads alongside other channels. That is not to say Google Analytics is irrelevant but measuring success purely on the methodology of clicks is a backwards thinking mindset in a world where digital advertising continues to innovate.