We’ve put more focus on LinkedIn in recent months, and for good reason! LinkedIn has made great strides in targeting options and functionality in the last 1-2 years since their acquisition by Microsoft. Those improvements make the platform more appealing to marketers which leads to more testing. More testing equates to more key learnings, which are exactly what I want to share with you today. We’ve identified some valuable audience targeting insights that might help you cut your eCPV for video campaigns in half!
The Opportunity
LinkedIn’s platform simplicity is both frustrating and advantageous for innovation. LinkedIn is more limited in targeting options than FB and gives you less control over bids and bid modifiers than Google. While that can be frustrating, it makes analysis easier and faster! With fewer variables to analyze, insights become more obvious. Insights are great starting points for new hypotheses and future tests, leading to more innovation down the road!
The Campaign Background
We launched multiple video campaigns, targeting sets of job titles, for a B2B brand. We chose job title targeting because we wanted the accuracy to be VERY precise. The service we are marketing to other marketers is very niche. However, I learned from A.J. Wilcox’s Hero Conf presentation that while job title targeting is very precise, it’s also the most expensive LinkedIn targeting option. After running the campaigns for 1-2 weeks, we began analyzing the early results in an effort to identify opportunities to lower the cost/view.
The Analysis
As I mentioned, the number of levers we can pull in LinkedIn feels limited, relative to Google and Facebook. In this case, I believe the lack of variables actually made our key finding more obvious to me. There are only two main areas where you can look for trends in your LinkedIn campaign data:
- The metric columns, which are grouped under the header “View”
- The demographic data that appears when you select one or multiple campaigns
There are eight demographics you can review when looking for trends which you can see in the snapshot below:
After analyzing the demographic data for our campaigns, we found no major trends in the data EXCEPT:
- The majority of clicks were from 3 seniority levels
- The majority of clicks were from the Marketing job function
The Hypothesis
If we can hone in on those particular seniorities within Marketing, rather than using job title targeting, we can reach roughly the same viewers and get cheaper video views.
How did we come to this hypothesis?
I also learned from AJ Wilcox’s LinkedIn Advertising presentation at Hero Conf that Job Function + Seniority is $ (lower cost) compared to job title targeting which is $$$ (highest cost).
Generally, I would have guessed that we could NOT get the same video engagement because job function + seniority is a broader group than the specific job titles that we knew were most interested in the B2B marketing service we were advertising.
However, when it came down to implementing the test, we made one targeting addition that seems to have made the difference.
We created the audience and as expected, job function + seniority targeting was very broad. In fact, it was larger than the audience size we wanted.
To narrow that audience, I added four unique skills as targets which qualify the type of marketers. These skills were narrowly focused, not simply “marketing” or “social media”. After adding those four skills, the audience size was just right. Our final campaign test included seniority, job function, and skills targeting.
The Results
eCPV was LESS THAN HALF of the job title campaigns!
More importantly, we lost NO quality in the views. The View Rates are still ~25% and the video completion rates are still ~23%.
The Conclusion
Use job title targeting as your first audience. Then, find where your video engagement and/or clicks are coming from based on the other demographics like seniority and job function. Combine those demographics with skills that you expect your audience to have, ditch the job titles, and lower your cost/view!
I hope this case study at least gets you digging around your audience data in LinkedIn. There are likely a ton of valuable insights for future testing!