Why is there a huge discrepancy between Google Analytics and Facebook ads?
We are running some Facebook campaigns that lead to our site i.e. "click external link" is the goal. The number clicks shown in Facebook ad manager are many more than those shown in Google Analytics. Is Facebook inflating the numbers? Or does Google Analytics just not capture every click?
Here is some more info I got from a Facebook support rep:
About the discrepancy that you found with your third-party reporting tool, there's something else I might add, even though, once again, since I don't have access to it I cannot see which configuration you have set for it:
- Facebook ad results will tell you how many clicks your ads have got. While clicking on your ad these users were still on Facebook - and that's why the metric is reliable, since they are clicking on a content that is still part of the Facebook platform. The service you are paying for - and the optimization on ad delivery, which is called Traffic - is made to let you get clicks on your ads, not visits on the website per se.
- Other third party reporting tools - such as Google Analytics - won't track clicks on your ad. Third-party reporting tools do not have access to your post Insights, they are not able to see how many likes, comments or clicks your boosted post have got.
These third-party reporting tools typically track visits on your website, not clicks. Users usually visit your website only after clicking on your ad. If you registered a high discrepancy between clicks and website visits you are experiencing what we usually call a high bouncing rate. This means that users did click on your ad but then they did not stay enough time on your website in order to be tracked by your third-party reporting tool. They might have clicked on the ad and then closed the website right away.
High bouncing rates are something that nowadays are affecting many Websites on the Internet and it's a complex situation that requires a deep investigation on the behaviors that users usually have while being on your website. A high bouncing rate might depend essentially from two different factors:
- Low internet speed, poor connection quality, old and slow smartphone models. 40% of mobile users close the browser webpage after 3 seconds of delay: if the webpage is closed before getting loaded completely, Facebook will still count the click on your ad but your third-party reporting tool won't track the visit.
Please consider that you were targeting India in these Ad sets. Unfortunately some parts of the world do not have a fast internet connection yet. A suggestion that I might give to you is to track people that is using a Wi-Fi connection or at least a 4G mobile internet bandwidth. You can also exclude older smartphone models in this very same section of the Detailed Targeting. You can add or exclude these targeting options in your Ad set by looking for Behavioral Targeting:
- Website that is not optimized for mobile phones, low or poor quality websites. As we usually say, the world has gone mobile - more than the 90% of the reach for these two campaigns went on the Facebook Mobile News Feed. Please be sure that your website is properly optimized for mobile users and that its content is appealing and relevant to your targeted audience.
The Relevance score for these ads have been of 1 or 2 on average, which means that the targeted audience did not react very well to your ad. You might need to change your creatives or to slightly change your targeted audience.
The Canvas is pre-loaded in the mobile phone memory, therefore it loads almost
instantaneously.In fact, the users that click on the button in the Canvas in order to go on your website will already have read more information than the ones that are clicking only after reading your post.
- Proper usage of UTM tags. I checked your ads and I could see that apparently you are not using tags that would make the sponsored URL unique. This means that when you select a general '
facebook.com' domain in your third-party reporting tool, you won't only see the visits that came from your ad. If you want to use your third-party reporting tool properly, you will definitely need to use them.
One thing to consider: Link Clicks vs "Clicks" For example if you have a boosted post that contains a video, any time people click 'play' in the video it will count as a 'click' even though they didn't go to your website.
This is true for all kinds of posts, even just boosted posts. If you have a regular ad for example, and the text is very long, facebook may have a 'read more' button to expand out the full text. When people click on the 'read more' that's considered a "click" too.
You can customize the columns that you see in reporting, so make sure you're looking at link clicks when you're thinking about clicks to your website.
Another key point that Sarah hinted at: Link Click vs Website Sessions/Users/Pageviews Google Analytics will only count someone when they actually load the GA script on your website. There are a few common reasons I've seen why this would happen: 1) people mis-clicked your ad. If you're advertising on mobile, especially "audience network" people may be clicking your ad unintentionally. Facebook shows their ads on many popular apps and games for android and ios. If you've played any of these you know they're absolutely covered in ads. Some games and apps may event try to trick users into click the ad to continue playing the game. Facebook is actively trying to shut down these abuses, but they still happen all the time. To fix this, simple remove the 'audience network' from your target. Most casual marketers shouldn't be using audience network unless they really know how to convert on that traffic. 'Mobile news feed' and 'desktop news feed' are the highest quality placements'.
2) bounces that GA misses If a user clicks on your ad on mobile, and your website loads very slowly, they might give up on your website before it's completely loaded and hit 'back' or close the website. Mobile users are using facebook with unreliable connections which means your website needs to load super fast to work on their device. This is technically considered a 'bounce' but GA will miss it because the user never got a chance to load the GA script on your website in the first place. To fix this, make sure to improve your page load speed so that the site loads completely in under 2 or 3 seconds on mobile.