Independent SEO Consultancy
She's in the Attic
She's in the Attic

Archaeology  ×  Literature  ×  Astrology  ×  Science & Art

[ Contact ]
Home× Astrology× Google Isn't Changing How Search Works
August 2026 Technical SEO Reporting Measurement

Google Isn't Changing How Search Works.
It's Changing How SEO Reporting Works.

Five separate changes in eighteen months. None of them touched rankings, relevance or retrieval. All of them touched the layer underneath — how rank trackers, SERP APIs, schema validators and analytics platforms observe what Google is already doing. If your reporting has felt slightly off lately, this is probably why.

None of these individually broke anyone's rankings. That's exactly why nobody's stringing them together. Five separate changes, all landing within about eighteen months, mostly covered in isolation as one-off SEO news stories. Not one of them touches ranking, relevance or retrieval — what actually determines who shows up where. What they all touch is the layer underneath that: how rank trackers, SERP APIs, schema validators and analytics platforms observe and measure what Google is already doing.

If your search reporting has felt slightly off lately, or a client has asked why a number moved that shouldn't have, this is probably why.

What's happening

JSON-LD parsing tightened August 2026

Google's crawler now applies a single pass of HTML unescaping to structured data, instead of unrolling entities repeatedly. Double-escaped entities (&) no longer get auto-corrected — they stay broken, and that can quietly kill rich result eligibility on a whole template at once, with no Search Console error to warn you.

num=100 parameter removed September 2025

No more single-request, 100-result scrapes of a SERP. Anyone pulling ranking data at scale — rank trackers, SEO platforms, in-house scripts — now has to make far more requests to see the same picture they used to get in one.

JavaScript rendering effectively mandatory January 2026

Dubbed the "SERPocalypse" at the time. Raw HTML requests increasingly return empty or degraded responses. If your tooling — or your own crawler — isn't rendering JavaScript, it's not seeing what's actually there anymore.

srsltid Shopping attribution tagging Ongoing from 2024

Been bleeding into ordinary organic (non-Shopping) URLs since 2024 and still isn't fully resolved. Product pages ranking in organic search get silently relabelled as "Organic Shopping" traffic in GA4 and Merchant Center reporting, skewing attribution for anyone running Shopping campaigns alongside SEO.

google.com/goto redirects Confirmed 26 August 2026

Organic result links no longer point straight at the destination page. They route through an encoded Google passthrough URL first, recoverable only by following the click — not by reading the results page. First spotted in testing in June, confirmed at near-full rollout by late August.

Why it matters

Your rankings haven't moved. Your real traffic isn't changing. Nothing here is an algorithm update. What's changing is the cost and reliability of everything that reads Google's results on your behalf — rank trackers, SERP APIs, schema validators, structured data pipelines, AI-visibility monitoring tools. Google is steadily raising the price of scraping its own output on a roughly quarterly cadence, whatever the stated reason on any given month — abuse prevention, ongoing litigation with scraping-adjacent companies, or genuinely just tidying up legacy behaviour.

"The risk isn't that your site gets penalised. It's that the dashboard someone glances at on a Monday morning starts quietly lying to them — and nobody notices until a client asks a question the data can't answer."

There's a second layer worth naming. As AI-driven search and answer platforms become a bigger share of how people find things, the reliability of the data feeding that measurement matters just as much as classic rank tracking — arguably more, since it's newer, less standardised and even easier to get quietly wrong. A JSON-LD parsing change that breaks a rich result silently is exactly the kind of thing that also degrades how well an AI crawler can understand and cite a page, since both depend on structured data being genuinely well-formed, not just visually fine to a human reader.

What to check now

A short list, not a strategy

Grep your live, rendered <script type="application/ld+json"> blocks for double-escaped entities (&amp;amp;, &amp;#) — check the actual output a browser or crawler sees, not the template source. The double-escape is usually injected at the theme or output layer, not written into the template itself.
If a rank-tracking dashboard shows a sudden swing with no matching change in Search Console impressions or average position, treat it as a collection artefact first, not a ranking change. Check GA4 organic sessions and actual conversions before reacting to any third-party tool.
Check whether product pages ranking in organic are showing up mislabelled as Shopping traffic in GA4 — the fix is a Merchant Center auto-tagging setting most people don't know exists, let alone that it's on by default.
Confirm which of your tools — rank trackers, SERP monitoring, AI-visibility checkers — are actually adapting to the goto redirect versus quietly returning gappier, laggier data without telling you.
If you're building or maintaining any AI-visibility or GEO monitoring setup, factor in that the underlying SERP data it depends on is now measurably harder and more expensive to collect reliably at scale. Worth knowing whether your provider has actually adapted or is just absorbing the gaps silently.

None of this needs panic. It needs someone actually watching for it, quarter to quarter — because the tools and dashboards you trusted last year aren't guaranteed to be telling you the truth this year, and the pace of these changes isn't slowing down.