The 'About' and bio depth experiment that moved less than expected

The question: how much does enriching About pages and author bios — the most-repeated EEAT advice — actually move rankings when tested in isolation?

A refreshingly negative case study, worth surfacing precisely because it underperformed. A B2B publisher added detailed author bios (credentials, sameAs links, photo, contact) to ~200 articles and rebuilt its About page with team, editorial policy, and physical address. Over four months: average position changed by ~0.4, inside normal volatility. No detectable lift.

This is consistent with what the QRG actually says: bios help a rater *find and verify* reputation that exists elsewhere. They do not manufacture reputation. If the off-site footprint is thin, a richer bio describes a thin entity in more words.

Counter-evidence: four months may be too short, and the bios could contribute to long-run trust calibration that this window did not capture.

Caveat: a null result from one publisher is not evidence of no effect generally — it is evidence that on this site, bios alone moved nothing measurable.

What we still don't know: whether bios act as a multiplier on existing authority (no authority to multiply, no effect) rather than an additive signal. This case is at least consistent with the multiplier reading.

Claude Cowork, Claude Design объединили в один Claude

➡️ Читайте на сайте: https://aff.top/blog/claude-cowork-claude-design-obedinili-v-odin-claude

🧠 Ещё больше инсайтов → в канале AFF.top

Microsoft планирует вставлять рекламу в игры

➡️ Читайте на сайте: https://aff.top/blog/microsoft-planiruet-vstavliat-reklamu-v-igry

🧠 Ещё больше инсайтов → в канале AFF.top

One author, 12 podcast appearances, and a citation-graph effect

The question: do non-link reputation signals — podcast appearances, interviews, mentions without backlinks — correlate with an author's entity strength and ranking?

A documented personal case: a specialist did 12 podcast appearances over ten months, most with show-note mentions but no follow links. No on-site changes ran in parallel. Reported result: the author's name began returning a richer entity result, AI Overviews started citing the author by name on topic queries, and the author's owned articles rose an average of ~1.8 positions.

This is interesting precisely because most mentions were nofollow or plain text — the QRG's reputation research (section 2.6) explicitly includes 'what independent sources say,' not just links. It is reputation without classic link equity.

Counter-evidence: podcasts that mention a person often produce secondary coverage, social shares, and the occasional followed link downstream, so a pure 'links don't matter' reading is unsafe.

Caveat: 1.8 positions is small and within the range of background update noise across ten months. The entity-result and AI-citation changes are the more convincing signals here, and both are hard to quantify rigorously.

What we still don't know: whether unlinked mentions feed entity understanding directly, or whether their downstream linked coverage does the actual work. The mechanism remains under-determined.

Building a product review that meets the reviews-system bar

The question: Google runs a dedicated reviews system favoring in-depth, evidence-backed reviews — what concrete elements satisfy its stated criteria for a single review?

Google's reviews-system guidance is unusually specific about what it wants to see. Built into a checklist:

— Demonstrate first-hand use through original evidence — your own photographs, measurements, or test results, the kind a spec-sheet rewrite cannot produce.
— Show expertise: convey knowledge of the product category and how this item compares within it.
— Quantify where you can. "Battery lasted 9 hours under continuous video" is the evidence the system asks for; "good battery" is not.
— Cover trade-offs honestly, including who should not buy it. One-sided praise reads as promotional, not evaluative.
— Compare to specific alternatives by name, so the reader can place your verdict.
— Explain how the product evolved from prior versions where relevant — the guidance names this explicitly.

Caveat: meeting these criteria improves a review's standing within the reviews system, but it operates alongside the broader quality and reputation signals. A rigorous review on a domain with a poor independent reputation is still constrained by that reputation, and affiliate intent does not exempt a page from the evidence bar.

What we still don't know: how the reviews system weights original media and measurement against textual depth, and whether it can algorithmically verify that media is genuinely first-hand rather than borrowed — a gap that surface-level signals could exploit.

Noindexing thin pages vs. rewriting them: a trust-signal decision

The question: when a section of a site is weak, is it better to noindex it or to invest in rewriting it to standard?

Google's guidance on the helpful-content system stated that unhelpful content on a site can affect how the whole site is assessed — framing low-quality pages as a site-level liability, not merely individually inert. That reframes the comparison: the choice is not 'rank this page or not' but 'does this page drag the site's overall trust assessment.'

Noindex (or removal) is the fast, low-effort move: it withdraws the page from evaluation. Rewriting is costly but converts a liability into an asset. The decision should turn on whether the page can plausibly become genuinely useful. A doorway-grade page with no reason to exist should go; a thin-but-legitimate page on a real subtopic should be rebuilt.

Caveat: Google has at times stated that noindexed pages are still crawled and that the site-quality assessment is more nuanced than 'prune everything.' Aggressive pruning has produced both recoveries and null results in publicly documented cases — the evidence is mixed and confounded by everything else those sites changed.

What we still don't know: whether noindexing fully neutralizes a page's drag on site-level assessment, or whether the system still 'sees' the underlying thin content during crawl.

A post-incident playbook when a published page damaged trust

The question: a page on your site turned out to be wrong on a consequential point — what response sequence repairs trust rather than quietly burying the error?

The QRG treats accuracy and accountability as core to trustworthiness, and how a site handles its own errors is itself reputation evidence. The post-incident playbook:

— Correct the substance first and fully; a partial fix that leaves related errors standing compounds the trust failure.
— Be transparent about the correction where the stakes warrant it — a visible correction note signals accountability, which the QRG associates with trustworthy sites, especially on YMYL topics.
— Trace the root cause. A single bad fact and a broken editorial process are different problems; the second recurs until the process changes.
— Audit adjacent pages for the same error class. One surfaced mistake often indicates a category of mistakes produced the same way.
— Strengthen the control that failed: the fact-check step, the reviewer assignment, the source standard — so the incident is a process fix, not just a content patch.

Caveat: transparent corrections aid trust but are not a documented ranking signal, and over-flagging minor edits as 'corrections' dilutes the signal. The case rests on accountability and accuracy as QRG trust components, not on a measured correction-note boost.

What we still don't know: whether algorithmic systems detect a site's correction behavior at all, or whether responsible error-handling reaches ranking only through the human-reputation channel and the reader trust it preserves.

FAQ schema markup vs. genuinely answering questions in the body

The question: to win question-intent visibility, is the lever FAQ structured data, or substantively answering the questions within well-structured body content?

This comparison shifted under our feet. Google reduced FAQ rich-result eligibility in 2023 to a narrow set of authoritative sites, removing the visual SERP payoff for most publishers. That collapses the case for FAQ schema as a traffic lever for the typical site — the markup remains valid but rarely renders.

Meanwhile, substantively answering questions — clear question-shaped headings, concise direct answers, supporting depth — is exactly what feeds featured snippets, People Also Ask, and generative-engine extraction. Studies of AI Overview and snippet sourcing through 2024 consistently find that pages giving a direct, self-contained answer near the question get pulled into summaries; the schema is not the mechanism, the answerability is.

So the honest reading: write the real FAQ for extraction and comprehension; deploy the schema only where you qualify for the rich result, and never as a substitute for actually answering.

Caveat: structured data can still aid machine parsing even without a rich result, so 'FAQ schema is dead' overstates it — it is just no longer a visual-SERP lever for most.

What we still don't know: how much generative engines lean on FAQPage markup for extraction versus simply parsing the prose, given that they reliably extract from unmarked text too.

An Author Bio Is a Pointer, Not a Signal
The question: does adding an author bio raise rankings? The advice circulates as a checklist item — attach a byline, a headshot, a paragraph of credentials, watch quality rise. The evidence for a direct effect is thin.

Bios are a surface feature. The QRG instructs raters to investigate reputation, but it directs them off the page — to independent sources about the author or site, not to the self-written bio. Per the reputation sections, a creator's own description carries little weight precisely because it is self-asserted; raters are told to seek what others say.

So the bio's value is conditional, not automatic:
— It helps when it lets a rater (or reader) verify a real, reputable entity that exists independently.
— It does nothing when the named author has no external footprint.
— It can backfire when credentials are exaggerated and contradicted outside the page.

Counter-evidence: sites that added structured author pages sometimes report ranking gains. Caveat: those rollouts usually coincided with content overhauls, editorial review, and genuine expert recruitment. Attributing the lift to the bio box alone confuses the visible change with the substantive one — a classic correlation-causation trap.

The defensible reading: a bio is a pointer, not a signal. Its worth equals the verifiable reputation it points to. A bio with nothing behind it is decoration.

What we still don't know: whether any current ranking system parses on-page author markup at all, or whether author-level reputation is assessed entirely through entity understanding built from sources beyond the page.