Cheap clicks, blind spots: what monday.com learned from ChatGPT ads

Clicks cost less. Signups took more work.

Yoav reported CPC at roughly a quarter of Google’s. Tamar said a smaller proportion of landing-page visitors became signups.

Those observations need to sit together. A cheaper click doesn’t tell you whether the customer was cheaper to acquire.

Someone asking ChatGPT to help organize a project might be relevant to monday.com without being ready to buy software. They may want to finish the work inside the conversation. An ad has to give them a reason to interrupt it.

The team therefore looks beyond clicks and signups to retention and paid conversion. They also described performance changing from month to month. This is still an experiment.

Their lean landing pages lost to pages with more content

The team found that richer, more informative landing pages worked better than the stripped-back pages they were used to testing.

That makes sense when you consider where visitors are coming from. Someone asking for advice may still need an explanation before they need a signup form.

For some campaigns, monday.com sent people to blog content rather than directly to a conversion page. The destination followed the conversation’s intent.

“Plan this project” beat another claim about being the best

Yoav described stronger results from ads that opened with an action, such as planning a project, rather than a familiar “best project management software” pitch.

The distinction shaped their campaigns. They separated people looking for a tool from those trying to complete a task or asking for advice.

Each situation called for different copy and a different landing page. Moving existing Google campaigns across wasn’t enough.

Conversion tracking was painful enough to mention twice

Tamar’s advice to new advertisers was emphatic: send conversion events back to OpenAI.

Getting that working was less smooth than on established advertising platforms. But she described a substantial improvement after moving from impression or click-based campaigns to optimization against conversion events.

That is a fairly unglamorous early lesson: before producing another batch of creative, make sure the platform receives the signals that tell it what a useful visitor looks like.

Their biggest frustration was spending without seeing enough

Asked about her biggest challenge, Tamar immediately named visibility.

They could see results, but lacked the search-level detail advertisers expect from Google. Yoav described a second problem: they couldn’t tell which relevant conversations they were missing, how much demand existed, or what share they captured.

That leaves a difficult question behind every campaign: is this a small opportunity, or are we reaching only a small part of it?

They’re also testing broader impact through geographic experiments. A person might see an ad, continue chatting, and search for monday.com later. Last-click reporting won’t capture that whole journey.

The question I’d add before bidding

At Aiso, we use a unique panel of 10 million real AI-human conversations to help founders and marketers understand where their brands appear and where opportunities exist.

One question deserves particular attention: does ChatGPT already recommend your brand in the conversations you want to pay to reach?

An organic recommendation doesn’t automatically make advertising wasteful. It does mean you should test whether the paid placement brings additional customers before scaling the spend.

Tamar and Yoav didn’t recommend Aiso during the episode. I checked. A disappointing omission, but the visibility problem they described is exactly what we work on.