A search consultant argues Google now grades a page by how it is built, not only by what it says. Koray Tuğberk Gübür laid out the case in a July 14 Search Engine Land column. He anchors it to a large unit-conversion site where one layout change preceded a sharp jump in clicks and impressions. The change: pulling a calculator tool up from low on the page to the very top, so it became what Google engineers call the page’s centerpiece annotation, the primary content block a document is judged by.
On that site, total clicks rose from 3.47 million to 4.53 million, a 30.5% increase. Impressions rose from 84.1 million to 167 million, up 98.6%, according to the case study Gübür cites. Average position improved only slightly, from 8.9 to 8.5. Average click-through rate actually fell 34.1%. That is worth sitting with: the gain came mostly from showing up in more searches, not from jumping far higher within them.
Gübür connects that result to a cluster of Google patents describing how machines parse structured information cards, product comparisons, and centerpiece annotation. That terminology entered public view through Google’s DOJ antitrust filings. He names Google researchers Michael Bendersky and Marc Najork, both tied to the company’s Gemini and generative-search work, as inventors listed on the relevant filings. Separately, Google’s own Search Quality Rater Guidelines list design effort as one marker of the human effort a page demonstrates, alongside accuracy and expertise.
None of that confirms layout is a scored ranking factor today. A patent describes what Google could build, not what its live systems weigh. Companies routinely file for ideas they never ship. The rater guidelines steer the humans who train and audit Google’s ranking models. They are not the algorithm itself, and design effort sits as one item on a long checklist, not a documented scoring weight.
The case study carries its own limits, too. It covers one commodity-answer site, a converter tool competing against thousands of near-identical pages for queries such as unit conversions. That site alone spans more than 100,000 programmatic URLs. At that scale, a single template edit gets multiplied across every page at once, which can inflate the visible swing in a way a smaller site would not reproduce. A single case plus a set of patents is a pattern worth watching, not evidence of a new ranking signal.
What survives the skepticism is the operating question, not the percentages. Gübür’s broader claim is that Google’s helpful-content systems weigh whether a page functions (lets a user calculate, compare, filter, or book) as much as whether it merely reads well. That lines up with Google’s own spam policy addition targeting “misleading functionality.” It is a lower-stakes, better-supported claim than a specific click lift tied to one site.
Teams with an underused tool, an unused comparison table, or a calculator buried below the fold have a cheap experiment available. Move the functional element above the fold before touching another sentence of copy. Then watch Search Console for impressions and clicks to move separately, since this case study suggests they can. Do not treat the 30.5% figure as a benchmark to expect elsewhere: it describes one high-volume site in an unusually commoditized niche, not a general algorithm change.
Reporting and analysis by Koray Tuğberk Gübür, published in Search Engine Land on July 14, 2026.
