SEO's Foundation Is Eroding, and the Steadiest Metric Left Is Whether Revenue Moves

Search used to reward patience. A site that earned its rankings could hold them long enough to build a reporting rhythm around keyword positions and traffic. That stability is gone. Google, Bing, and the AI answer engines now retrain and reweight often enough that a site's own citations can fall in the same week its off-site mentions rise, and the dashboards marketers built their credibility on no longer describe anything steady. The operators still getting traction have stopped defending the old scoreboard and started proving the one number that survives the volatility: revenue.

Jake "The Wizard" Tlapek started in U.S. Air Force IT before spending two decades in marketing, building and selling two agencies along the way. The second, Wizard Marketing, was acquired by Finch, where he served as chief evangelist until his full exit this summer. He now works independently and runs Business Arcanum, an application-only peer group for founders and marketers. In his last conversation with Agentic Marketing News, he argued that messaging outlasts execution; this time his focus is the search layer itself, and his read is that it's harder to stand on than it's been in years.

"The more personalized we get with AI, the less power SEO is going to have, and the more the prioritization is going to be on the person," says Tlapek. "Our messaging, all of our SEO, has to be consumer-alignment focused. We have to align with who they are and their personality, not feature lists."

The floor keeps moving

Tlapek's read on the current search environment is that it has lost the consistency that made it measurable. The engines that decide who gets surfaced, Google and Bing and now ChatGPT among them, are changing their models fast enough that the same optimization can perform differently from one day to the next. "It's really, really hard to create an analytically consistent SEO program right now," he says.

He points to a case he saw recently, an operator whose on-site citations were falling while every mention of the business hosted somewhere else kept climbing. Transparency, in his description, is at a low, and results can swing by the day, which leaves agencies defending outcomes they can no longer explain with the reports clients are used to reading.

Tie the story to revenue

His prescription starts with moving the reporting conversation downstream, toward tying spend to outcomes rather than the predictor metrics the industry has trained clients to watch. Keyword counts and ranking positions were always proxies, and Tlapek argues they're now unreliable enough that leaning on them puts an agency's credibility at the mercy of a volatile system.

"The only metric that matters is revenue," he says. "Is revenue changing? Is revenue improving?" The obstacle is behavioral. Marketers spent 15 years teaching business owners to love those proxy numbers, and owners are slow to unlearn them. "My favorite thing to say about business owners is that their feelings are always right," he says. "How to fix them usually isn't." The same discipline shapes how he thinks about channels, where he'd rather see a marketer put a full year of real effort into one than spread thin across five.

Owned data and off-site proof

With every marketer drawing from the same handful of models, Tlapek's guard against the resulting sameness returns to a point he's made before, that the work starts with a strong messaging plan built on benefits rather than features and on the emotional driver underneath a purchase. Paired with it is data a brand actually owns. Proprietary stats, original studies, and metrics no competitor can cite give an operation something unique to put into the world, and that uniqueness is what earns weight where a model has to choose.

The other half of his current playbook lives off the brand's own site. Reviews are carrying outsized influence in how AI systems assemble a shortlist, and so are organic mentions in the places buyers already gather. "SEO is so much less about your own site, your owned media," he says. Getting named in Reddit threads, Quora answers, and industry forums, and turning up in comparisons without paying for the placement, is the kind of mention off your site where he sees a lot of the current advantage, and it rewards the judgment a machine lacks about which rooms are worth being in.

Selling to one person

Tlapek expects the shift to accelerate as buyers hand more of their comparison work to AI. He puts agentic purchasing six to 12 months out, the point at which a shopper tells an assistant to buy a vacuum and the assistant places the order, and he thinks the more personalized those systems get, the less leverage traditional SEO holds. What replaces it is alignment with the individual. Before his last exit he was building product pages that changed as they loaded, matching a visitor to one of 54 captured personas and adjusting for whether that person values price, style, or brand.

His insurance example makes the logic plain: a family photo, or even the house in the background, could shift to resemble the visitor's own life, pulling on emotional cues a buyer registers without noticing. "We're going to be writing content to sell to a single person," he says. As shoppers keep outsourcing the comparison to their assistants, the work he's betting on is consumer-alignment work: build around who the buyer is, because the feature list is already the machine's job.