I often get people asking questions like:
Why do the Google Analytics conversions attributed to Google Ads not match up with the counts in Google Ads
Yesterdays Google Ads conversion counts are way too low. Your tracking is broken
Google Analytics is not reporting all my conversions.
This article clarifies how different systems deal with conversions and why they will never match up.
The main focus is on the difference between Google Ads and Analytics conversions, but the same principles apply to other advertising and analytics systems.
- Analytics picks who gets the credit for a conversion
- Ads take the credit whenever they can
- Analytics reports at the time of conversion
- Ads reports conversions at the time the ad was shown
- How they track users and sessions may vary
- Do they track users as they switch between devices?
- Are they blocked by cookie consent options?
- Are they blocked by ad blockers?
- Do they track repeat conversions?
- How far back do they look to determine who gets the credit?
- The only conversion tracker you can trust is the one that processes the conversion
Who gets credit for a conversion?
Analytics systems give credit to one
Analytics systems will attribute credit for each conversion. I’ll explain how Google Analytics Universal does it, as most systems will do something similar.
With Google Analytics Universal, it will determine that the credit is given based on a “last non-direct click” attribution model. That is, what was the last known system that sent the user to the website before they converted.
A direct click means the analytics system could not determine where the user came from. As this is of little use, Google Analytics will ignore it and look back to see if it can give credit based on how the user had previously visited.
“Last click” means Google Analytics only looks back to the last way a user visited the site that was not direct. All previous visits are ignored.
If all visits from a user were classed as direct, then direct will get credit for the conversion.
Google Analytics defines who gets the credit using the ‘source’ and ‘medium’ dimensions. Each visit (session) from a user has the source and medium set. If the user converts, then the last source/medium visit that is not set to direct is attributed to the conversion.
You can further refine credit into Channel Groupings. These are built on a set of rules which often use the source/medium values to decide which channel each conversion is placed.
Google Analytics has a “top conversion path” report to help clarify this. We can look at the source/medium for the most popular paths a user goes through before converting.
The next diagram shows some relevant user paths with the source/medium that get the last non-direct click attribution highlighted in green. “google / cpc” represents Google Ads. The source was Google, and the medium was Cost Per Click, which means an advert. For this article, it is important to note that two of these examples attribute Google Ads (google / cpc) with the credit for the conversion. Two are attributed to Google My Business, and one is unknown, so marked as Direct.
I was not 100% truthful about analytics systems giving all the credit to one. Some systems do support alternate attribution models that can distribute conversion credit around. Google Analytics does have reports like the one above to help you dig into alternate attribution models.
Google Analytics Four now supports a cross-channel data-driven attribution model. And this means it’s even less likely that each system will attribute conversions in the same way.
Google Ads takes all the credit it can
This is where we find out the first significant reason why Google Ads conversion numbers will not match the numbers in Analytics. Google Ads will take credit for every conversion it was involved in. Not just last-click, but any click. With our previous examples, Google Ads takes credit for three of the conversions, not two.
Note that all advertising channels will do the same, which means multiple advertising channels may take credit for the same conversions, and the numbers will not add up.
When we are talking about the attribution models in Google Ads, we are talking about how it internally attributes conversions to ads. A very different thing to how analytics systems attribute the system that caused a conversion.
Some analytics and ads systems like Google Analytics and Google Ads provide a way to track a user as they move between different domains that you control. This way you can avoid attributing a conversion to another one of your domains. When this is set up, movement between the domains is not considered a touchpoint in the attribution model, and therefore can’t take credit for a conversion. If different systems do cross-domain tracking differently then you will see differences in the user’s path for each system and therefore how attribution is done.
What time is a conversion recorded?
Analytics systems say it as it is
It’s true. As expected, a conversion is reported at the time it happened.
Google Ads does some time travelling
Google Ads reporting conversions on the date of the ad impression, whereas many other reporting tools attribute them to the date of the conversionGoogle Ads Help
Google Ads reporting and data gathering is based on ad impressions. How often was an ad shown, was it clicked on, did it convert. This makes sense as you want to know how well specific ads perform, and the smart algorithms need that data to work out which type of ad impression works best.
Google Ads reports conversions at the time of the ad impressions that caused the conversion, and not the time of the conversion itself. This means conversions can be reported on different days to when the conversion actually happened. And another reason why you can’t try and compare numbers with analytics.
A conversion may be minutes, days, or even weeks after an ad impression. To report accurately, Google Ads would need to travel into the future to know if an ad impression is going to cause a conversion!
They can’t do that. And that is why they give a warning when looking at data from the last few days. In this case, they expect most conversions to happen within four days.
Please don’t read the recent conversion numbers as gospel; they may change. I followed a single day in Google Ads and took daily snapshots of the conversions reported on that day. In 8 days, the conversion count went from 2 to 7.55! Do not look at recent day conversion counts in Ads.
7.55 conversions, how can that happen?
Fractional conversions happen when you use a conversion attribution model that distributes the credit for conversions between multiple ad impressions. e.g. a linear attribution model will give all ad impressions an equal share of a conversion. Let’s go back to the user path diagram in relation to Google Ads.
The second path has the user visit the site twice due to Google Ads. i.e. they saw and clicked on two ads. A linear attribution model would give each ad impression 0.5 conversions, while a last-click model would give the second ad impression the full conversion value of 1.
The third example has the user click on four ads (and they first found the site via a Google search 🙄), so each ad impression would get 0.25 conversions credited to them.
Using a non last-click attribution model is another reason why numbers will not match the Google Analytics reports, which are last-click based.
I’ll not go into detail on this, but each tracking system works differently. They may determine visits/sessions differently, they may identify and track a user in different ways (some can track users between devices), they may identify where a user came from in different ways. All this will mean variations in the resulting data used for reporting.
Users may block tracking in different ways. From using ad blockers to filling in different cookie consent option. Expect each tracking system to lose a certain percentage of data and therefore conversions. Don’t be surprised if this reaches 20% or a lot more if you are asking for consent. And it will get worse.
If the server handing the conversion does not see a conversion, then the conversion never happened.
Server-side tracking is less vulnerable to being blocked, especially if the server runs your site and handles conversions. If you want accurate conversion reporting, don’t look at third party data like Google Analytics; look at the data coming from the system that runs your site and handles the conversions.
Sometimes a user may convert multiple times. i.e. make two purchases. Some systems may record both while others just the first one. Google Ads lets you configure if an ad click can have more than one conversion in the conversions settings.
Systems will look back for a certain length of time with regard to user activity, and what they did beyond that time is ignored. Google Ads, by default, looks back 30 days when attributing a conversion to an ad impression. Any ad impressions beyond 30 days do not get any credit for the conversion.
Only use your Google Ads reports for optimising Google Ads. The same goes with any of your advertising channels.
Use analytics reports to give you a way to compare how your channels are performing.
If you want an accurate count of conversions, only trust the horse’s mouth, the system that handles your conversions.