When running a social media campaign for a B2B client, the first platform you probably think of is LinkedIn. LinkedIn provides the most premium audience in B2B. No other social media platform allows you to target decision makers by job title, company, and industry. Your second thought may be Facebook (and Instagram). These channels, on the other hand, are less tightly targeted for B2B but still hit the right people if the strategy behind the targeting is tactical. With Facebook providing greater reach at a more efficient CPM, and LinkedIn providing a tighter target with higher CPMs, the platforms often complement each other when combined in a larger social media strategy.
In deciding on a platform (or platforms), we don’t typically think about bots, but we should. Bots have been an issue since digital advertising started. Many platforms have been very transparent in their efforts to combat bots, with the goal being to cut down on wasting spend and misleading campaign results. But how effective are those efforts really?
We analyzed one of our B2B client campaigns and noticed a lot of traffic coming through LinkedIn based on its in-platform metrics, but we saw a large drop-off in users when looking at Google Analytics (GA). Often ~45% of LinkedIn clicks to the landing page result in a GA session, with one instance dropping as low as 6%. Facebook, on the other hand, was consistently above 90%. In order to understand what was going on, we implemented a test with FullStory, which provides session replays and error monitoring for each user who made it to the site. As it easily integrates with GA, we were able to record all of the interactions and study trends of users coming from LinkedIn and Facebook individually and compare the results.
As many would expect, there was a difference in user interaction based on where the user saw the ad, but the results were more staggering than expected. We found that there was a very large percentage of traffic coming from Linux OS. This accounted for 44% of the LinkedIn traffic, while Linux overall accounts for 1.85% of all operating systems in the United States (source). According to this article from Adobe, bots will often times use Linux as their operating system.
FullStory assigns each user a unique name and will show how many sessions a user has had in their lifetime. Out of all 245 Linux sessions, there were no repeat users, despite many of the session times coming within minutes of each other. According to the same article by Adobe, this is a sign of bot traffic as “Bots usually get a new visitor ID every time they execute, thus incurring only one visit ever and all their traffic will consist of a visit number of one.”
Additionally, all 245 Linux users have almost an identical pattern. They all have a session count of one, with the total session seconds all being between four and five seconds, and an active time of zero seconds, except for three users, which all have a time of one second. Every Linux session also matched back to one of two recorded latitudes and longitudes. Click here to see the chart.
For comparison, Facebook only has 0.21% of its traffic coming from Linux, and those have active session times of 27, 24, and 13 seconds, which are much more realistic.
We also found that 71% of the traffic coming from LinkedIn has an average active session time of zero seconds. Facebook is the opposite, with 63% of the sessions averaging a time of greater than 10 seconds. Click here to see the chart.
So, what does this all mean?
As of now, bots are making up about 44% of the traffic in our LinkedIn campaign. Every other platform will block bots and credit money back each month for any that do sneak through. LinkedIn seems to be behind on bot blocking and crediting back any spend that occurs due to them being so strong on their platform. We cannot conclude that this is always the case, but advertisers should be cautious when running within LinkedIn and leverage software that verifies the site users beyond LinkedIn’s in-platform metrics.
LinkedIn, arguably, has the best targeting for B2B clients, but there is a strong possibility that you may spend media dollars on bots and inflated metrics while running within the platform. There are still cases when, even with these bot results, LinkedIn can be a good platform for your media strategy, but it is important to understand what you as an advertiser are getting for your media dollars. Is LinkedIn’s premium audience worth it?
Andrew Tilley is a manager at Brunner with a focus on all things paid social. Andrew has worked with clients of all sizes from startups to Fortune 100 companies as well as both B2B and B2C brands in a wide variety of industries. He holds certifications across major social platforms including Facebook/Instagram, LinkedIn, and TikTok.