This, seemingly an easy question, drags us into a very deep rabbit hole, and it’s called attribution models. Have you heard of it? If yes, then you know a thing or two. If not, you’re in for a big surprise.
I’m literally sitting down, writing these words, and I still don’t know where to start. OK, focus, deep breaths and let’s go.
In this post, I will explore different attribution models and try to explain how Google Analytics and Google Ads count conversions. This will shed some light on why your GA4 account and Google Ads conversion don’t match. Or any other ad account, like Facebook.
I will also show you where to change how conversions are counted and if this is something you should do.
What are attribution models?
Attribution models have been around for a while. I think they launched at least ten years ago in Google Analytics. But why? Well, this goes back to how users behave online. Most users’ journeys are not linear, meaning that they might come to your site many times; some may use several channels, and some only one. They also have different consideration times. If the product is expensive they might take their time to think, ask their wife or husband and come back after days, weeks or even months.
This behavior is natural, but was hard to track by analytics platforms. Up until recently Google Analytics reported last click. This means that whatever the channel user came from right before converting, got all the credit for driving that conversion. Is it fair?
Many companies invest money in different channels and they also do email marketing. How to distribute that credit equally amongst all those channels? And most importantly are there channels, that don’t contribute to a conversion at all.
That’s why different models were introduced. I won’t mention them because Google Analytics decided to sunset several models. We will have only data-driven and last click models available.
This all might be confusing, so let’s look at an example.
Let’s say a user click on your Google ad, came to your website and purchased, or converted.
This is a simple scenario, and in this case all the credit for that conversion goes to Google Ads. We obviously see that the user came through the ad, to the website and on to become a customer.
Let’s make it a bit more complicated. In this scenario user left the website without converting. And then came back later by clicking on your Facebook ad. And this time they converted.
Who gets the credit in this situation? We can’t say it was all Google Ads, as user left and came back via Facebook ad. Would the user purchased without first visiting from Google Ads? Maybe. And we can’t give 100% credit to Facebook ad, as user was already familiar with the website from previous visit. Maybe he had time to think about his purchase?
This is why you see different numbers in your Google Ads account and Google Analytics. Google Ads uses last ad click attribution model and Google Analytics uses data-driven attribution model. In both platforms this can be changed, I will get into that later.
In this scenario Google Ads will take the credit (conversions in the account) because it was the first ad click. What happen later, doesn’t matter.
Google Analytics uses data-driven approach where it looks not only at channels, but at devices, time and many other factors to determine who get’s the credit. So the numbers might not be the same. As Analytics might attribute conversion to Facebook ad. In some cases they might match, but a lot depends on your marketing mix.
How to change attribution model in GA4?
Analyzing attribution in Google Analytics is not a part of this post. But I will show you how to change it in case you want to play around.
Go to your Google Analytics account and click on Admin. It’s on left bottom corner. Find Data display and click Attribution settings.
There you should see a drop down that allows you to change your attribution model.
Easy, right. Here’s what you should know:
- Changing the attribution model will apply to both historical and future data. This is cool!
- changes will be reflected in all reports that use event-scoped traffic dimensions (Source, Medium, Campaign, and Default channel group).
- You’ll notice changes to the following metrics when used with event-scoped traffic dimensions: Conversions, Total revenue, Purchase revenue, and Total ad revenue.
- You may see decimals (or, “fractional credit”) for the first time in these columns. This is because credit for a given conversion is distributed between contributing ad interactions according to your selected attribution model. This is important as some people get confuse when the see 1.5 conversions (one and a half).
But should you change anything at all?
You can compare the two models in your GA account and decide if one of them shows a better picture. Find advertising in your left side menu, last icon. Then select model comparison.
You can compare those two models and see which channels are undervalued or overvalued. This is how it looks for one of the e-commerce sites and worked a while back.
For the first two channels the data-driven model shows more conversions, which means, that they probably contributed to more conversions than just from their last click. There might have been users who came from those channels, did not converted and probably came back through other channels.
But there are also channels for which the data-driven model shows less conversions. Those channels might be closers, as I call them. Most likely users don’t need to think a lot they just come and convert. It could be ads promoting deals, it can be email with discount offers.
So in this case if I switch my default data-driven model to last click, I might not see any difference. But what I can draw from this is find the channels that are closers and double down on that, knowing they are good at converting users and they are undervalued in GA account (default reports).
You don’t have to switch models if you don’t see a big difference. But if you do switch closely monitor your actual sales. Not what you see in the account, but what you see in the bank. If the change is big you should see more revenue coming in after a while as you double down on channels that were underreported before.
Also, don’t switch attribution models in the middle of a high season, like Black Friday, Christmas, summer or winter. Try to find a season where your conversions flatten out.
If you scroll down a bit, you will see setting: Channels that can receive credit.
This is where you can tell Google Analytics if you want the Google Ads conversion credit to be share with Organic channels as well. What it means is that if you select the first setting the credit will be given only to Google Paid channels. Even if the user visited other channels after Google Ads click and then converted, the credit will go only to Google Ads.
On the other hand, the second selection will distribute conversion credit between other channels if user visited them before converting.
Which setting to choose again depends on how many marketing channels you use. For example, if you have a pop up on your site offering a discount for the first purchase, email might play a big part in converting the user. With the first setting, email channel will be ignored since user came from ads. In the second setting the email will get part of the credit as well, since it was in the users path towards a purchase.
Conversion window
This is often overlooked setting. A conversion widow determines how far back to look to attribute a credit for conversion. For example, Google Ads has a default conversion window of 30 days. If a users visits the site on day 1 and leaves. Then after 15 days, the user comes back (no matter the channel) and converts, Google Ads will attribute that conversion to an ad click on that first day.
That is also the reason you might see different conversion numbers in GA or any other platform. That setting can be changed, both in Google Ads and Analytics. But you should know why exactly are you changing it.
The easiest way to do it is go to your Attribution settings in Google Ads and in the left side menu find Path metrics.
If you see that most of the users convert within the first day or two, you can reduce your conversion window to 7 days. Don’t forget to do a “common sense” test. If this data shows you a certain amount of days, look at your product and ask yourself if that is true. For example, for a cheap product 1 day to convert makes sense. But if you have a $500 course, or service or $1000 watch, most likely users will need more time and visits to purchase your product.
So make sure the data aligns with how people would actually buy your product. If yes, change it and monitor.
Why you should change it?
If you reduce the conversion window, less conversion will be attributed to Google Ads. Which means less conversions in the account, in your campaigns and keywords. This is good because you might see that you have been spending money on conversions that would have happened anyway. In other words, you were overspending.
Other channels will get more credit as well. This will show you that maybe you were undervaluing some of the other efforts. But proceed carefully.
You might also want to increase your conversion window. What if your product is so expensive that people need couple of months. For example, a car or travel arrangements which might be quite expensive and be planned in advance. Some people book summer travels in spring or winter. They might have click the ad in January, but after some thinking and exploration they might have purchased everything only in March. Then the keyword and the ad that actually brought those people to your site, won’t get any credit at all. And you might pause that keyword loosing sales in the long run.
Don’t changed anything if you don’t have data to support that change.
How to change attribution model in Google Ads?
In Google Ads you have to go into Tools & Settings in the top menu and find Attribution under Measurement column. You can also compare the two models last-click and data-driven. Find Model comparison on the left menu.
You can select only a few campaigns or all. And don’t forget to select one conversion from the drop down. Looking at all conversion might be confusing. Just like in GA, look at campaigns, their conversions and value and see it it’s worthwhile to change the attribution model in Google Ads as well.
You might wonder what’s the difference in changing attribution here and in GA? Well, GA shows all your traffic and conversions along with different channels. This is only Google Ads specific. And most importantly this will be used in your bidding strategies.
If you see that changing models will bring you more conversions, on your left menu click Switch to DDA and check if your conversion is eligible for a switch.
Select the conversion your want to switch and proceed. You will get this informational pop up. I suggest you read it. Here or in your account.
I’m confused, why should I do this at all?
I would be surprised if you weren’t. It’s a lot to take in, especially if this is the first time you’re hearing about it. Honestly, you don’t have to do anything. This is advanced stuff and you should explore it only if you have done everything else. Or there is something wrong with the account and you tried everything and the data shows that the model change might help.
At the start of the post I wrote that the goal is to show you why the conversion numbers are different. This does not mean you have to do something about it.
Even if you import GA conversions into Google Ads you might see different numbers. I mentioned in the earlier post that I prefer Google Ads conversion pixel, with a few exceptions.
Keep the default settings if you’re unsure.
Attribution models: key points
- Attribution models are tools used to determine how credit for conversions is assigned across various marketing channels. They have evolved due to the non-linear nature of customer journeys online.
- Google Ads and Google Analytics use different attribution models, leading to discrepancies in reported conversion numbers. Google Ads employs the “last ad click” model, while Google Analytics uses a “data-driven” approach.
- You can change attribution model in Google Ads and Google Analytics. But this should be done if you have data to support this change.
- The change in attribution models affects both historical and future data across various metrics like conversions and revenue. You might start seeing fractional credits for conversions.
- Compare both attribution models to identify underreported or overvalued channels, and adjust marketing strategies accordingly.
- Don’t change attribution models frequently, especially during peak seasons. If possible, make the change during periods of stable conversions for accurate assessment.
- Check your conversion window, and if data aligns with your business, change it and monitor the results.