This past winter, I helped an organization launch their first campaign fully tagged and tracked in Google Analytics. This included tagging emails, posts on Facebook, LinkedIn, Pinterest, select tweets on Twitter, and for the first time experimenting with ads and boosting posts on Facebook and LinkedIn. Using Google Analytics’ standard reporting for traffic leads from each source, we began to compare efficacy of each medium and adjust efforts put into the campaign. For the ads however, while the reported CTR was excellent at eventually 2.39% (for the three-month campaign), in comparing reports from Google Analytics with reports from Facebook, we realized that the amount of reported clicks by Facebook was greater by nearly an order of two. I calculated what the real Facebook ad CTR was from the reported impressions / visits in GA, and found a saddening .092%. What accounted for this huge discrepancy?
At first, my mind jumped to deceit. I had watched a pretty compelling experiment done by the YouTube channel Veritasium this February, pointing out the elephant in the room of pages with bloated amounts of ‘likes.’ I am glad to say that I discovered the discrepancy, and while this experiment should be seriously considered in spending large budgets on Facebook (ala General Motors in 2012), the issue ended up being much more benign.
As it turns out, both Facebook and LinkedIn factor in social actions into their CTR calculation. After revisiting specifically the boosted posts (instead of the ads in the sidebar), I quickly realized what was driving the clicks. In most of our posts, we clearly displayed a link, however in one post in particular the link was removed and all that was left was the image. This image was not just pulled from the social graph, but was a literal image that was posted from our social media team. And what else to do when there is an image with a description? Click the photo to get the full-screen, theater-box view of the image, of course.
12,533 people (8.01%) didn’t visit our site, unfortunately. 12,533 people saw a giant image and clicked on it, but 24 people clicked on our link that we provided. We realized that in our workflow, the only data immediately given to me was the CTR straight from Facebook’s report, but after visiting the Facebook page myself the other reported information seemed like a minor oversight to not clearly show in their exported-Excel-sheet report.
So what were the lessons learned from this discovery?
- First of all, if you want engagement, use images… just like everyone says.
- If you would like site traffic, clearly display a separated link for users to click on. Nothing intrusive, but intentionally clear.
- The farther away you get from the source, the harder it is to know the truth. An integrated, well-connected team is never a bad thing.