Facebook analytics reliablity issues you need to know
| Justin Kistner
Being the most visited website in the world, Facebook naturally uses a CDN to keep up with traffic demands. CDNs boost page loading times by providing local copies of the Facebook servers near a user’s location. This means that Facebook aggressively caches their images. Several Facebook analytics solutions that John Lovett reviewed are dependent on loading a single pixel image. I thought I’d shed some light on why Webtrends does not use images to tag Facebook fan pages, apps, or tabs.
Cachebreakers
Since Facebook aggressively caches images, using a standard image tag does not provide an accurate count for your analytics. As a hack to get around this limitation, some vendors are attempting to break the cache by adding a random string to image source, which is more commonly known as cachebreaking.
The Facebook developer wiki has a section dedicated to cachebreaking that says, “In short, please don’t append these “cachebreaker” strings to the end of your image source strings.” They then provide 3 reasons why this method is a bad practice:
- Since each user’s profile has a separate copy of that image, you actually increase the likelihood that some of those images don’t get cached successfully.
- Not having the image cached successfully is bad because if the new image does not get cached successfully, then the user will be left with a blank spot where the image should be. In other words, if the image doesn’t load, then the analytics count is not accurate.
- And finally they say “It causes Facebook to potentially cache many, possibly millions of, copies of the same image, which fills our image cache, increases network traffic between our servers and yours as we refetch the image each time, and generally hinders our ability to help serve you better. ” That means Facebook needs to prevent this kind of activity in order to provide a good user experience to its members.
Part of the reason we don’t use cachebreaking methods is because it increases the amount of times that an image won’t load and that provides unreliable data. We also didn’t feel comfortable putting our clients at risk of Facebook cracking down on cachebreaking.
FBJS Libraries
Other Facebook analytics solutions rely on FBJS libraries. The problem is that Facebook does not allow FBJS to run when the custom tabs load. A user has to interact with the custom tab in order for FBJS to run. To get around this, vendors use image tags and cachebreaking methods, which again are a bad practice.
To provide reliable Facebook analytics in compliance with the Facebook Terms of Service, Webtrends has developed a patent-pending method for data collection. We leverage our data collection API, which is a service included with our analytics software. I invite you to join me on April 7th for a webinar to learn more about how Facebook marketing works and how you can track it.