July 28, 2021
Industry Insight
Curated Team

Testing the sh*t out of LinkedIn ads, so you don’t have to

When it comes to testing a new social channel or trying to get the most out of the one you are already using, it often takes a test and learns approach. Well, if LinkedIn ads are on your list of channels to test, let us help you out by sharing everything we learned when we tested the sh*t out of LinkedIn for one of our clients.

Now, we’d used LinkedIn’s ad platform before for a number of clients, to great success. But, in all honesty, we pretty much felt we knew what worked and stuck with that. And it did work.

However, like everything social, things change quickly, algorithms change, ad platforms introduce new features, and audience consumption changes too.

Enter our latest LinkedIn brief from a B2B client: test everything you can on LinkedIn and tell us what works. 

Establishing Goals

Goal one was to establish the best strategy for the client going forward, based on the results of the expansive tests we would run. But, ultimately, there were goals that underpinned this — namely, to drive leads among a very niche targeting a pool of IT specialists. 

At the end of this project, we needed to have established the best methods on LinkedIn for driving these qualified leads: what ad type works best for this audience, what’s the best way to reach them, do they engage better with video or image ads, are they willing to share their details for good content, what messaging resonates with them? 

Challenging Assumptions

As mentioned, having run many LinkedIn ad campaigns previously, we had certain assumptions about what would work and what wouldn’t. While we approached this testing project with an open mind, it’s fair to say we thought we knew what many of the outcomes would be.

To put it bluntly, my money was on a pretty bog-standard approach of sponsored content traffic ads, using either still images or videos, leading to a landing page with gated assets. 

We also anticipated a reach and retarget approach, building awareness with ungated content and then remarketing to site visitors with gated assets — we tested both written whitepaper content as well as event sign-ups.

These were fair assumptions based on what had worked for us previously. And they weren’t entirely wrong. But, in testing everything, we expanded our horizons to the possibilities of what could be effective on LinkedIn. 

Everything We Tested

I’m not suggesting this was an entirely exhaustive test of the LinkedIn Campaign Manager platform — perhaps you’ve tried other approaches beyond what we included — but we did a fairly good job of giving everything a try. Not to simplify our approach — believe me, it was not a simple process — but we kind of threw everything at the wall to see what stuck!

Here’s a rundown of what we tested:


  • Website visits / conversions
  • Lead generation

Ad formats

  • Sponsored content: single image ads and video
  • Within this, we tested a wide variety of creatives — images with people and without, those with text and without, different videos, etc. 
  • Text ads
  • Message ads (previously InMail Ads)


  • By job title
  • By job function
  • By member groups
  • Company list uploads

All of these were layered with extensive exclusions to refine our target audience.


  • Whitepapers
  • Webinars
  • Virtual roundtable events


  • Long copy
  • Short, snappy copy
  • Messaging with statistics and figures
  • Various CTAs

Bid optimisation 

  • Manual bidding
  • Automated bidding 

As you can see above, our tests included elements offered by the LinkedIn platforms themselves — ad type, audience targeting — as well as elements more within our control — creative assets and messaging. We also ran all these tests in various combinations with one another; video ads with lead gen forms, text ads with copy variations, message ads with different content offerings (webinar, roundtables, etc.).

All in all, we ran a lot of ads over a few months, keeping track of every success and failure, making optimisations along the way. And we learned a lot. 

What We Learned

Within our control

Firstly, as it turned out in this case, image type and messaging had less to do with performance success than we might have first thought. At least that was the case for this audience and this goal. 


While it’s true that copy with standout stats sometimes had the edge, there were factors that seemed more important in determining success. I’d always still recommend testing copy variations in your ads — in fact, LinkedIn themselves suggest running at least 4 ad copy variations within each of your campaigns — but the ad type and objectives you choose are far more likely to pack the biggest punches. 


Before we jump into ad formats and goals, it’s also worth stating that, while creative assets, like a copy, had less to do with determining success, there were still some learnings here. Still imagery trumps video. Now, of course, if you’re running top-of-funnel brand awareness campaigns and just want to get eyes on your brand, video awareness ads could be a great shout. But, for this lower-funnel, lead gen activity, we found the video to be more of a distraction to the ultimate goal. By watching a video, people feel like they have taken some form of action — however passive an action it may be. As a result, they seem to be less likely to take another action, the action you really want them to take. 

As such, we found greater success with still imagery for driving our ultimate lead goal. That said, the type of imagery was less of a sticking point — images with people, which we often find work well, performed no better than our imagery with text or generic stock images. Again, I’d always recommend testing what works for your campaigns, and your audiences, but the greater focus should be on ad formats, content offering, and goals.

Within the LinkedIn platform

Here was our biggest learning: native lead gen ads work. And they work well. 

Objectives and ad formats

Our initial assumption was that this highly targeted, the senior audience would need more nurture time. We planned to send them to the site, get them to engage with content, and then retarget them with a more meaty offering — a webinar or exclusive roundtable in exchange for their details.

And that’s what we did for a while. But we soon found that sending them to the site ended up being an extra unnecessary step in the process. Once we started testing lead gen ad formats, where LinkedIn users within our target audience could fill in the form without ever leaving LinkedIn — in fact, most of it was auto-filled, so they barely had to fill in anything — our lead conversion rate shot up. 

As it turns out, this super-busy, highly qualified audience actually needed less nurturing than we thought. If they are interested in the content you’re offering, they want it now, as simply and quickly as possible. 

We quickly traded our traffic-driving ads for lead generation across all our ad formats — including sponsored content and message ads. 

Message ads

Speaking of, message ads gave us another interesting learning. Firstly, no matter how refined your audience is, with all the best will in the world, sometimes your ads land in front of irrelevant people. Either their job title is ambiguous or they belong to a member group you might not have expected — either way, they don’t fit the bill for who you’re looking to target.

Nine times out of ten, if you’re running newsfeed ads, these irrelevant audience anomalies happily skip past your ad, understanding it’s probably not quite the right fit for them. Annoying, yes — for both you and them — but, impact on the overall campaign, minimal.

However, when these less relevant individuals get an ad sent directly to their InMail inbox, of course, it feels much more personalised. Perhaps their curiosity takes over even if they feel the content doesn’t quite align with their needs. And so you get a host of irrelevant — I’d go so far as to say ‘random’ — leads. This was our biggest sticking point with LinkedIn in this whole campaign — for all its benefits, it’s not perfect, and this is a clear example of its imperfections. 

Audience targeting

A quick note on audiences: if you speak to any LinkedIn rep, they will tell you the best way to reach your audience is through a combination of Job Function and Seniority. It makes sense on paper — job titles change, new quirky job titles are introduced, and someone who is highly relevant to your campaign may slip through the net if you are targeting solely on a specific set of job titles.

However — and if our LinkedIn rep happens to read this, I’m only half sorry — I respectfully disagree with this approach. We tested a variety of targeting methods: by specific job titles, by job function, and by relevant member groups. For this very niche campaign, where we were looking for very specific job titles, the job function angle just didn’t cut it. 

Here’s a tip: if you’re looking for reach, and specificity is less of an issue — member groups are a great way to boost your campaigns’ visibility. 

Post-test Thoughts

Now, of course, our test was finite, it only ran for a relatively short period, and since then, we have discovered some additional learnings. The main one being: lead gen ads are great — but not on their own. 

Over time, lead gen ads — like all ads — can run into ad fatigue. The way to combat this is through additional supporting activity. Since running this test project, we have made lead gen ads a focus of a large number of our B2B campaigns. But what we have found is that, without supporting activity, the lead volume declines over time. To counter this, we have begun running higher-funnel video ads to boost awareness among our audiences, alongside our lead gen approach. 

So What’s the Takeaway?

Here are my suggestions to you — keep testing, throw away the rulebook, and give lead gen ads a fair shot. 

Your audience likely won’t be the same as ours, your objectives may differ, your creative assets will be varied — so you need to test and see what works for you.

That said, given what we found, lead generation ads can be a great way to seamlessly generate leads within the platform, making the user journey for your potential customers as easy as possible. Give them a try!

And throw away your assumptions — your reach and retarget strategy may work, but is it blinding you to alternative approaches? Maybe your audience doesn't need as long a nurturing time as you think, maybe you’re overcomplicating the process. Or maybe not — but you won't know unless you try! 

Alternatively, if you too want to test the sh*t out of LinkedIn but don’t know where to start, we’ve got you covered. Get in touch and let’s discover together!

Our recent posts