Customer story
Built on a real Askalytics AI Insight generated against Poker.org's connected GA4 and Search Console data. The implementation steps and early outcomes are the team's own and are still being measured.
Poker.org ranked near the top for some of the biggest search terms in poker, on desktop. It just was not winning the clicks. One Askalytics AI Insight sized the leak, named the pages, and showed Samuel's team how to recover most of it without earning a single new ranking.
From one AI Insight against 60 days of GA4 and Search Console data.
US desktop, more than mobile, at 0.4% CTR.
Desktop clicks vs mobile, on more impressions.
Projected recovery if desktop matched mobile CTR.
Verdict: strengthened by a second AI.
Poker.org is the global poker community's home page: news, strategy, hand rankings. Samuel Hellman is the product director, and his job is the long game, the audience and traffic that hold up between viral news cycles.
Organic search was already the biggest channel. On paper it looked strong. But Samuel had a nagging sense the site was ranking better than it was performing, that it sat high in results and somehow still left traffic on the table. Proving it meant cross-referencing Search Console impressions against device and country, and that was never quite worth the afternoon it would take.
We were ranking for terms most publishers would kill for. The traffic didn't reflect it. I suspected a click-through problem but I couldn't point at where.
A Traffic Acquisition AI Insight came back for Samuel. It came back at a confidence score of 92, the highest in his queue, with a Confidence Check verdict of "strengthened." The headline was blunt: "U.S. desktop SEO is the clearest traffic leak: 3,626,274 impressions at 0.4% CTR vs U.S. mobile's 2% CTR."
The finding was the thing Samuel had felt but never proven. US desktop generated 3,626,274 impressions but only 13,550 clicks at a 0.4% click-through rate. US mobile, on fewer impressions at 2,112,920, pulled 42,612 clicks at 2%. Same site, same rankings, five times the click rate on mobile. The insight did the math: if desktop matched mobile's CTR, the same impression base would imply roughly 72,525 clicks, about 58,975 more than the site was getting.
It was honest about the limits too. GA4 showed zero tracked ecommerce revenue across channels, so the insight framed the prize as audience growth, not a tracked sales figure, and flagged that conversion tracking should be fixed before anyone tried to value the clicks in pounds.
It put a number on a hunch. Nearly 59,000 clicks we already rank for and aren't getting, and it told me which pages were the worst offenders. That's a quarter's worth of growth from titles we already own.
The insight ranked three actions:
1. Rewrite desktop titles and meta descriptions on the worst offenders. It named them: the poker-hands-ranking-chart page at 2.9M impressions and 1.1% CTR, the beginners-guide-to-the-rules-of-poker page at 444,720 impressions and 0.4% CTR, and the best-poker-chips-set guide at 0.2% CTR.
2. Use US mobile's 2% CTR as the benchmark for the desktop work, since that is what the same content already earns on another device.
3. Add or validate non-ecommerce value tracking before judging monetization, because GA4 reported zero revenue and conversions across every channel.
It handed me the exact URLs and the exact benchmark. My editorial team didn't need a second meeting to know what to test first.
The title rewrites were the obvious quick win. The team reworked the desktop titles and meta descriptions on the three named pages first, leaning on the framing that already earned 2% on mobile. That was an afternoon of editorial work per page, and it shipped inside the first week.
The benchmark task was simple to honor: they set US mobile CTR as the target and built a small weekly Search Console pull to watch desktop CTR move toward it.
The tracking task was the hard one, and they could not do it the way the plan suggested. Poker.org runs no ecommerce, so there was no purchase event to wire up, and a full "value per engaged reader" model meant changes to the ad stack that sat with a different team on a longer timeline. Rather than stall the whole project waiting on that, the team took a different route: they treated recovered organic clicks against their known display RPM as a rough proxy for value, enough to prioritize and report on, and parked the proper tracking model for later.
We couldn't build the full value-tracking model the insight wanted, so we used RPM as a stand-in. It was good enough to act on. The title work was the part that actually mattered, and that was easy.
These are early-measured and projected numbers, not a finished experiment. In the weeks after the title rewrites, desktop CTR on the three named pages started climbing off the 0.4% floor toward the mobile benchmark. The full opportunity the insight modeled was roughly 58,975 recovered organic clicks. Even capturing a third of that is close to 10,000 extra readers a month on content the site already ranks for.
Put against the team's display RPM as a rough proxy, recovering that traffic is worth far more than a year of Askalytics costs. The strongest part of the story is the leverage: no new content, no new rankings, just titles rewritten on pages the site already owns. One insight pointed the team at the single highest-return SEO work available to them.
Recoverable with zero new rankings.
Titles rewritten on the worst offenders.
This is the rare SEO win where you don't have to earn anything new. The rankings were already ours. We were just leaving the clicks on the table.
The Bloom & Measure demo property is loaded by default on every free trial account.
Samuel suspected a click-through problem. The insight sized it at ~58,975 clicks and split it cleanly by device. Ask AI let him drill into the worst pages himself.
Three exact URLs with their impressions and CTR, ranked by opportunity. The detectors pull those figures straight from Search Console, so there was nothing to second-guess.
With no ecommerce revenue tracked, the insight refused to invent a pound figure and flagged the tracking gap instead. That restraint is why the team trusted the rest.
The thing that won me over was where it said "I can't size this in revenue yet." A tool that admits a limit is one I'll act on.
Platform: Askalytics. Data: GA4 and Search Console (read-only). See also: channel optimization, how every number gets checked, and why this beats raw GA4.
Connect GA4 and Search Console in a few minutes. Your first AI Audit runs the same day.