If your LinkedIn targeting is already tight and CPA is creeping up, Conversation Ads can be the quiet leak: the format looks “higher intent,” but the delivery signal may be worse than the simpler alternative.
That’s the uncomfortable part. Message-style ads are often benchmarked with eye-popping engagement—open rates commonly cited in the 50–60% range and CTR around ~12% in some benchmark summaries—far above typical Sponsored Content CTR benchmarks like 0.44% (sources: Search Results summaries in the Research Brief). Yet AJ Wilcox’s April 2026 write-up argues a head-to-head test can flip the story: Conversation Ads can open less and cost more than plain Message Ads, even when everything else is held constant.
And because LinkedIn often sits at the center of B2B paid social—41% of paid social budgets in one 2023-oriented expert summary (Research Brief)—format mistakes don’t stay small. They compound into pipeline math.
That’s also where Verto fits. If your team is trying to connect ad-format decisions to actual pipeline outcomes—not just platform engagement—Verto gives you a cleaner way to see what’s happening across the funnel. Instead of treating opens and clicks as the whole story, you can use Verto to pressure-test whether a format is really producing qualified meetings, opportunities, and revenue, which is exactly the kind of visibility you need when a “higher intent” ad unit may be quietly underperforming.
The contrarian finding: “more features” can mean fewer opens
Wilcox’s core claim isn’t that message ads don’t work. It’s narrower—and more actionable: Conversation Ads (the branching, multi-CTA variant) can under-deliver compared to Message Ads, even when they’re designed to look identical.
“We’re seeing Conversation Ads come in around a 35% open rate, while Message Ads are closer to 50–55%, with everything else being the same.” — AJ Wilcox, “Stop Using LinkedIn Conversation Ads (Here’s What to Use Instead),” published Apr 27, 2026
Same audience. Same sender. Same subject line. Same copy. Same CTA. Different outcome. That’s the PaidLab problem in its purest form: a targeting constraint can be perfect, a creative can be fine, and the format still changes the physics of delivery.
But the data doesn’t stop at opens. Wilcox ties the drop to unit economics:
“When you run the exact same campaign, Message Ads are coming in at about 40% lower cost per lead compared to Conversation Ads.” — AJ Wilcox, Apr 27, 2026
That’s not “optimize your funnel” advice. It’s a post-mortem. The leading indicator (open rate) degrades, then the lagging indicator (CPL) follows.
To understand why this matters now, it helps to zoom out: LinkedIn has spent the last few cycles adding more testing and governance inside Campaign Manager (built-in A/B testing, expanded duplication options, work email validation for Lead Gen Forms—Research Brief). The platform is basically telling operators, “run cleaner experiments.” So do that. Especially when a format-level issue might be real and persistent.
What to use instead: one primary move (and the trade-off)
If you only change one thing, change this: swap Conversation Ads back to Message Ads, and rebuild the “retargeting you’ll lose” with a simple holdout-based sequence using LinkedIn’s other native levers.
Wilcox calls out the main reason advertisers switched in the first place: LinkedIn lets you retarget users who engage with Conversation Ads, but not those who engage with Message Ads. That’s a real feature gap. It’s also a trap if the top of the sequence is opening materially less.
Trade-off (say it out loud): this will reduce retargeting convenience before it improves efficiency. The “engagers” pool may shrink. The point is to buy back opens and CPL, then recreate downstream intent with formats that have better control surfaces for ops teams.
Seen from the other side, the “instead” is not mystical. It’s just cleaner qualification:
- Message Ads for the direct outreach moment (optimize for opens and clicks).
- Lead Gen Forms when you need structured conversion and lead-quality governance (work email validation was added in 2023—Research Brief).
- Document Ads when the offer itself does the filtering (the Research Brief includes a case example: 30+ quality leads at ~$60/lead for a BIM software client in a tight local market).
- Thought Leader Ads when credibility is the bottleneck (employee posts cited as “8x more reach and engagement” than brand posts in a 2023-oriented summary—Research Brief).
This is the creative qualification lens: pick the format that forces the prospect to “pay” with the right kind of intent (time, form friction, or social proof), not the one that gives the advertiser the most branching CTAs.
Run it this week: an Experiment Lab you can actually read out
Here’s the 5-minute version you can run this week—without pretending the platform dashboard proves incrementality.
Hypothesis (make it falsifiable): If we replace Conversation Ads with Message Ads to the same matched audience, then open rate will increase and cost per lead will decrease because the Message Ad format appears to deliver with higher inbox visibility (as observed in Wilcox’s 2026 testing).
Setup:
- Audience: identical targeting for both arms (same company list, titles, exclusions, frequency caps). No mid-flight edits.
- Creative: same sender, subject, body copy, and a single CTA. (This is crucial—otherwise it’s not a format test.)
- Conversion path: keep it constant. Either both click to the same landing page, or both use Lead Gen Forms. Don’t mix.
- Experiment tool: use LinkedIn Campaign Manager’s built-in A/B testing capability (introduced in 2023—Research Brief) if available for your objective; otherwise run two campaigns with rigid naming and no shared budget optimization.
Launch:
- Budget range: set a fixed test budget per arm that’s meaningful for your account (directional, not definitive). The goal is to get stable open/click signals, not to “win” the quarter in seven days.
- Timeline: 5–10 business days, or until each arm reaches a pre-set minimum delivered volume.
- Owners: Demand Gen owns setup and creative parity; Marketing Ops owns naming conventions, lead routing checks, and readout integrity; RevOps confirms downstream definitions (MQL/SQL/meeting) before launch.
Readout: Success = CPL (or cost per meeting, if that’s your true north). Guardrails = open rate and CTR (leading indicators), plus lead quality checks (work email validation pass rate if using forms). Stop-loss = if one arm’s CPL is trending ~40% worse after sufficient delivery, pause it and document the delta (Wilcox’s 2026 claim gives a reasonable threshold to watch for, even if your account lands somewhere else).
Next test: If Message Ads win on CPL but you miss retargeting, rebuild the sequence with a two-step: Message Ad → Document Ad or Lead Gen Form retargeting based on site visits or form opens (depending on your instrumentation). Keep the measurement tied to meetings/opportunities where possible—mirroring the Understory case framing around meetings and closed deals (Research Brief).
When “stop using Conversation Ads” is wrong
Blanket bans are usually lazy. This one has a boundary.
Conversation Ads may still make sense when the retargeting pool is the product (for example, very narrow ABM lists where you’re deliberately buying engagement segmentation), and your own head-to-head data doesn’t show an open-rate penalty.
But if the goal is qualified pipeline efficiency, the operator move is to treat format choice like any other constraint: test it, isolate it, and don’t let a feature checklist override the unit economics.
The neat trick in Wilcox’s piece is that it isn’t really about a format. It’s about discipline. When an ad type “should” perform—because benchmarks say message ads can hit 50–60% opens and double-digit CTR (Research Brief)—and your account doesn’t, the fix isn’t more branching logic. It’s a cleaner experiment, a stricter creative control, and a willingness to pick the boring option when the numbers keep pointing the same way.
