# From 80k AI articles cut to 180: unanswered.io changed its shell just before Google's cleanup and its rankings kept climbing · Web

> Source: https://stringers.io/en/web/web_unanswered.io.html (Stringers / 什么值得做, generated 2026-11-13)
> Stringers tracks which small products suddenly took off each day, investigates why one by one, and distils playbooks from them. Numbers come from 15 databases we collect daily, each with source and date.
> For agents: these investigations and the underlying databases (verified revenue, App Store / Google Play charts, ad libraries, GitHub, Product Hunt…) are queryable live over MCP — tools feed (daily breakouts) / plays (playbooks) / hot / products / data. Setup: https://github.com/komako-workshop/stringers (the user signs in with Google at stringers.io/en/start.html to get a personal URL). Site index: https://stringers.io/llms.txt

> Earlier this year it used AI to publish 80k long-tail articles to capture English Google search traffic. In September, after Google cleaned up bulk content, it deleted thin pages on its own, rewrote about 180 winning pages as Q&A, and inherited the old URLs' authority. Seventy percent of its views now come from these pages.

[Web](https://stringers.io/en/web/index.md) · 11-10 written

Unanswered(unanswered.io,Zyref, LLC) · unanswered.io

# From 80k AI articles cut to 180: unanswered.io changed its shell just before Google's cleanup and its rankings kept climbing

**What it is** An English Q&A site where each question gets one answer, drafted with AI assistance, under a shared byline, with sources and a review date. Topics come from the questions people search for most on Google.

**Why it took off** Early this year it used AI to publish 80k long-tail articles and picked up English Google search traffic. In September Google cleaned up bulk content. The site proactively deleted thin pages and rewrote about 180 winners as Q&A, inheriting the old URLs' authority. Those pages now account for seventy percent of views.

**C** **How you could follow suit** Don't build a generic AI version of Quora. Pick a time-sensitive vertical close to money, seed volume based on search data, keep the winners and delete the rest based on real clicks, then rewrite in the format of "direct answer up top + sources + review date + public AI disclosure".

Grade C: the playbook of "seed volume, filter winners, swap the shell and move authority" is worth learning, but the English Q&A space is being squeezed by AI summaries overall. The site itself hasn't monetized and gets only a few thousand visits a day, and its traffic hangs on the November Google core update. Few people know about it yet

Global rank 255k · 90 days ago, top 1M+

unanswered.io looks like an English Q&A site that launched in late October: one answer per question, with a byline, sources, and a review date. In fact it has existed since early this year, starting as a content site where AI produced 81k articles in two months, drawing traffic from long-tail English Google searches. After Google's September crackdown on bulk content, it deleted nearly all 80k of them, kept the roughly 180 that people actually read, rewrote them as Q&A, and passed the old URLs' authority over. Its global rank climbed from 950k to 260k in 40 days. Below: how it operated over this year, where its traffic came from, where the money is, and whether you can copy it.

### The page someone landed on after searching for how to run Walmart delivery

"I want to use Spark Driver to deliver for Walmart. Each order shows just one total, and I can't tell how much is base pay and how much is tip, or whether the tip gets clawed back. How does Spark calculate pay, and how much can a driver actually make per hour?"

This is the most-viewed question on unanswered.io, posted Oct 29 20:32 by an account called `unanswered` that had signed up two seconds earlier. The answer opens with the verdict: "You get paid per order, not per hour." It then cites a driver-tracking app's stats on 50 million drivers' 2025 earnings, with a median of $21.74 an hour, and gives a table of how long customers can lower a tip (3 hours after delivery), raise it (24 hours), or add one later (14 days). It ends with the 2026 February settlement between Walmart and its drivers, $100M.

By Nov 11 this page had 8,315 views. It wasn't written on Oct 29: that day the site moved in 181 questions at once, each replacing an older article that people were reading and other sites were linking to (inference: created in one batch that day, with views two to three orders of magnitude above the newer content). These 181 pages took about 70% of the site's views since the redesign. The top ones were Walmart Spark driver pay, Facebook Protect, how to delete a Meta account, DoorDash earnings, Android pop-up virus, localhost refused to connect, and whether Instagram costs money.

These questions share a pattern: lots of people search them, the answers go stale ("what are the rules as of today"), and they sit close to money.

### What it is

- **Format:** an English Q&A site. Each question has one answer, signed "staff writer," with a list of sources, "reviewed / next review" dates, and a revision history. It also has daily quizzes, user questions, reputation points, and bounties, clearly modeled on Stack Exchange.

- **Who runs it:** Zyref, LLC, behind which is Pakistani developer Rehmat Alam. Since 2020 he has built a Google search results scraping API (SerpsBot) and a professional-social-data API. The latter was shut down in October 2025 after that social platform sued, and settled in September 2026. Neither his personal site nor the Zyref site mentions unanswered.io.

- **Is it AI-written?** The site says so itself. Its editorial standards page (updated 10-31) reads: "'Staff writer' is a byline shared by our editorial team... we use AI tools to assist with research and drafting." 13 staff writers (James Whitfield, Maya Okafor, and others) each have 40 to 120 answers credited to them, and all have 1 reputation on the site. Only two human moderators are visible.

- **Does it make money?** Not yet. The page has only Google Analytics, with no ad code, affiliate links, or paid membership.

### How much the ranking rose

| Date | Global website ranking (Tranco) |
|---|---|
| before end of September | Not in the top 1M |
| 10-01 | 952,849 |
| 10-11 | 595,634 |
| 10-20 | 449,716 |
| 10-31 | 300,906 |
| 11-03 | 287,037 |
| 11-10 | 258,649 |

This ranking is a 30-day rolling average. Working backward, real traffic began rising around mid-September, accelerated through October, and slowed in November. The site's own public view counter shows roughly 13k views over 103 days since the redesign, about 8 thousand a day, which matches the 260k-rank scale: a few thousand visits a day, not a big site.

Chrome's real-user rankings show an earlier timeline: India entered the top 1M in June, the US, Canada, Australia, and the Philippines entered the top 500k in August, the UK and Netherlands in September, and Poland and Vietnam in October. These are all English-speaking or English-heavy countries, which is what a long tail of English Google searches should look like.

For comparison, from August to November, Quora, answers.com, ask.com, reference.com, and wikiHow stayed flat or dropped. Sites that mass-produce FAQs at scale fell harder: faqtoids.com dropped from 140k to 470k, and questionsanswered.net from 270k to 550k. The whole category is being squeezed by Google's AI summaries, and unanswered.io is moving against that trend.

### 80k pages to 180: what it did over the year

| Timeline | What happened |
|---|---|
| Jan 31 | Registered the domain, with no prior history |
| Mar 4 | Homepage read "the directory of the internet's questions": 75,711 guides in 54 categories, claiming expert review and an average rating of 4.7 stars (per archived pages) |
| Apr 8 | 81,088 pages, then no more added until end of August |
| Sep 29 | Rebranded as "independent, sourced explainers," with author pages showing licensed-professional personas such as "Julian Moreau, LMFT" (marriage and family therapist) |
| Sep 24 to Oct 8 | Google's September anti-spam content update |
| Oct 28 to 29 | DNS moved to Cloudflare; new Q&A system launched, with 181 questions loaded at once |
| From Oct 29 | Old guides with traffic or backlinks redirected to new Q&A pages; the rest returned 410 (permanently deleted); all the licensed-persona author pages were 410 as well |
| From Nov 4 | 150 to 220 new Q&As published per day |
| Nov 10 | Google's November core update began rolling out |

On Nov 10, I opened five old guide URLs at random ("10-10-10 eye rules," "23 and AncestryDNA comparison," "incline bench press 30 or 45 degrees," and others). All of them were 410.

The whole operation runs on a script. Its public API docs include a publish endpoint described as "publish an editorial Q&A," with fields for the attributed staff writer, sources, whether it was verified, and the next review date, plus a `legacy_paths` field: "the old /guide/ URLs this page replaces; as long as they still appear in search results, the old URLs redirect here." So migrating old articles, redirecting them, and publishing two hundred new items a day all go through the same endpoint.

### Where users come from

**Almost all of it is organic Google search.** The evidence adds up: the topics are all long-tail English search queries, views cluster on pages that inherited old URLs, real users appear only in English-speaking countries, and the site itself says "we start from the questions readers search for most; these are marked Popular question" and "only editor-approved answers go into the sitemap for search engines."

**AI-answer citations leave some traces, but the volume is unclear.** Captions on an Instagram short video, the description of an Apple podcast, and a guitar forum post all contain text like "unanswered.io +1," which looks like someone copied it along with citation markers from a ChatGPT-style answer. But a third-party AI visibility report shows "no visibility" in the US.

**Almost no other channels.** No ads, no social accounts, no mentions on X, no Product Hunt or Hacker News launch, no press coverage. The site's community features just launched: of roughly 500 registered accounts, many are link-spam marketing accounts, the daily quiz leaderboard reads "No scores yet," and 420 questions have no answers.

### Why this one broke out

1. **It tested with volume first.** Two months and 81k pages covering high-search long-tail topics like income, bills, platform policies, and device problems, letting Google decide which ones ranked.

2. **It cleaned up before the crackdown.** In September, when Google went after mass-produced content, it didn't wait to be demoted. It deleted 80k thin pages and the author personas with fabricated licenses, concentrated its authority on about 180 winning pages, and adopted a format Google tends to favor: a direct answer up top, bylines, sources, review dates, revision history, and a public editorial policy that even discloses AI use.

3. **The people behind it know search data.** They spent six years building a Google search results scraping API, so they know what people search for and which keywords are valuable, and they can build an automated pipeline that publishes two hundred items a day.

4. **The topics are chosen where AI summaries struggle to replace them.** Walmart Spark tip changes that can be made within hours, when Facebook Protect must be turned on, how to cancel a Sam's Club membership: these questions need facts "as of a specific day," and answers with review dates have an edge.

To be clear: **one question per answer, sources, and review dates are not why it worked.** The traffic came from the 80k pages built before the redesign; the new format is there to protect that traffic and make it look compliant. Q&As published after the redesign get single digits to a few dozen views each.

### Who it's competing with

| Competitors | Global rank (11-10) | How it differs |
|---|---|---|
| Reddit | 109 | Discussion threads, which Google has ranked heavily over the last two years; unanswered's pitch is "no digging through threads" |
| Quora | 505 | Many answerers, uneven quality |
| Stack Exchange | 1,086 | Professional community; unanswered's reputation, accepted answers, and bounties are modeled on it |
| wikiHow | 2,077 | How-to tutorials, with real-world testing |
| answers.com / reference.com / ask.com | 4,600 / 6,200 / 7,600 | Established Q&A and content sites, large but not growing |
| faqtoids.com / questionsanswered.net | 470k / 550k | Mass-produce FAQs by search demand and monetize with ads; now being demoted by Google |
| Google AI Overviews / ChatGPT / Perplexity | — | They give the answer directly and are the biggest threat to this whole category |

This is a crowded, shrinking space. unanswered.io is much smaller than these established sites, and its only distinctions are "new" and "growing against the trend." I haven't found anyone copying the whole "AI mass-production → Q&A wrapper" playbook, but the craft of mass-producing content by search demand has been around for twenty years.

### Where the money is

Right now, zero. If it added display ads: 8 thousand views a day, about 240k a month, at $5 to $20 per thousand views, that's roughly $1,200 to $5,000 a month (estimate; a sizable share of its readers are in India and the Philippines, which pulls the rate down). The topics lean toward Medicare, tax filing, Verizon bills, and memberships, high-CPM areas, which suggests monetization was planned for later. It may also be aiming to become a "sourced citation" for AI or to funnel traffic to sister products; nothing outside shows this (unverified).

The biggest variable is coming up: **the Google core update rolling out from Nov 10 could knock it down directly.** Right now about 70% of its traffic sits on 180 inherited old pages, and the new content hasn't yet proven it can carry traffic on its own.

### If it were you

**Verdict: C. There's a clean playbook to learn from, but copying it into a general English Q&A site isn't worth it.** The category is being eaten by AI summaries, the site hasn't earned a cent yet, a few thousand visits a day can't support a business, and the whole thing hangs on Google's next update.

**Who it's for:** teams of 1 to 2 people who already run a content site, or who hold search data in a specific vertical and can write scripts and do SEO. Especially those sitting on a pile of half-dead old articles: the step of "keep the winners, delete the rest, consolidate authority into them" applies directly.

**Who it's not for:** anyone starting from scratch to build an "AI Quora"; anyone without search data who picks topics by gut feel; anyone who can't stomach losing half their traffic to one algorithm update.

**How to do it:** don't build a general Q&A. Pick a vertical that is time-sensitive and close to money, such as a gig platform's pay and rules, certain bills and memberships, or tax and benefits in a specific country. Choose topics from search-term data and publish at volume first. After two or three months, pick the winners by real clicks and delete the rest or keep them out of the index. Rewrite each winner in its format: a direct answer up top, a source list, review dates, revision history, and a public editorial policy that discloses AI use. Don't fabricate licensed authors. Build the publish endpoint with a "which old URLs this page replaces" field from day one so migration and redirects happen automatically. Only put reviewed pages in the sitemap.

**How to get traffic:** organic Google search, plus being cited in AI answers wherever possible. The site kept pages that were cited in platform forums and Facebook groups, which suggests backlinks remain the most reliable signal for deciding whether a page is worth keeping. Refresh pages regularly so the "as of today" date stays current.

**When to act:** the tactic has no hard window, but Google tightens with every update. If you want to try it, finish the first round of "publish at volume → filter" before Q1 of 2027, and see how much search traffic in your niche is still left after AI summaries take their share.

**What would show I'm wrong:** after this November core update finishes rolling out (usually two to three weeks), if unanswered.io's rankings rise instead of fall and new Q&As reach several hundred views each, it would mean this "AI drafting + editorial format" holds up in Google's eyes, the category is wider than I think, and the rating should go up. Conversely, if it falls back below 500k in December, it would show that a new wrapper can't save mass-produced content, and the lessons drawn from it lose value.

[Website ↗](https://unanswered.io)

**Research notes: competitor table, acquisition evidence, full research report**

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| Quora | Web | 2009 | Tranco 505 | Many-user UGC answers; unanswered: one question per answer, editorial bylines, sources, and review dates |
| Reddit / Reddit Answers | Web | 2005 / 2024 | Tranco 109 | Discussion threads; unanswered's pitch is no thread-digging, just one answer |
| answers.com / ask.com / reference.com | Web | Long-established | Tranco 4.6k / 7.6k / 6.2k | Established Q&A / content sites, large but not growing; unanswered is a low-cost AI-era copy |
| wikiHow | Web | 2005 | Tranco 2,077 | Mostly how-to tutorials with real-world testing; unanswered covers a wider range, leaning toward income, bills, and platform policy |
| Stack Exchange | Web | 2009 | Tranco 1,086 | Professional community; unanswered copied its reputation / accepted-answer / bounty mechanics, but the content is mostly published under a single shared editorial byline |
| faqtoids.com / questionsanswered.net / consumersearch.com | Web | Old guard (System1 / IAC family) | Tranco 470k / 550k / 130k, all trending down | Same playbook: batch-generated FAQs built around search demand, monetized with ads. unanswered has no ads and adds sources, a review date and an editorial standard |
| Brainly / Gauth | Web | 2009 / 2020 | Tranco 19,641 / 15,361 | Homework Q&A aimed at students |
| answerbag / blurtit / theanswerbank | Web | 2000s | Tranco 130k–250k | Old Q&A site, basically stalled |
| Google AI Overviews / ChatGPT / Perplexity | Big-company features | 2023–2024 | — | Gives the answer directly, the biggest threat to this category. unanswered's format is built to be cited by them |

Where users come from

- **Confirmed** **Google organic search (English long tail, programmatic SEO)** : 2026-03's homepage says 75,711+ guides, 81,088 articles in April; CrUX shows only English and near-English countries; 70% of site views land on pages migrated from the old /guide/ URLs 301 181; the site says topics come from search demand and only editor-approved pages go into the sitemap

- **Confirmed** **Old URL link-equity migration (301 / 410)** : the OpenAPI publish endpoint has a legacy_paths field. Of 26 old guides I checked, 4 301 to new Q&A pages (all previously linked from outside sites), and 22 return 410

- Inferred **AI answer citations** : found traces on Instagram, podcast show notes and forum posts where users paste an 'unanswered.io +1' citation chip. But seomonitor shows no AI visibility in the US; the recall card's claim about LLM citation index #869 is unverified

- Inferred **Community / daily answer emails** : about 500 registered accounts (many are link-spam accounts), the answer leaderboard has nobody on it, and 420 questions have no answers

- Inferred **Paid ads / social / press** : the page only has GA4 G-32EYGL54Q8 and no ad code; 0 mentions on X; no press, Product Hunt or Hacker News records

## unanswered.io: why it showed up

**In one line** : unanswered.io is an English Q&A site built by Pakistani developer Rehmat Alam through Zyref, LLC. Each question gets one answer, with sources and a "review date". It's not actually a new product. The domain was registered 2026-01-31. From 2 to April it batch-generated **81k "guides"** with AI, a programmatic SEO play that picked up real traffic from Google long-tail search. In late September Google rolled out a spam update, and it then **converted to a Q&A community** in 10-29. Of the 80k articles, only about 180 with traffic and backlinks were kept and rewritten as Q&A. 301 of those redirect to the new pages, and the rest return 410 (deleted). New content is published at 150–220 items a day under a "staff writer" byline (the site itself admits AI drafts it). The rise in Tranco rank mostly reflects Google long-tail traffic built up by the old guides. The community and daily-answer "retention" features only just launched, and almost nobody uses them.

### 1. Data: how much it grew

**Tranco** (traffic library, rank_day):

- It was not on the weekly chart at all through 07-30 to late September (not in the top 1M). It charted at #952,849 on 10-01 and climbed every day after that: 10-11 595,634 → 10-20 449,716 → 10-31 300,906 → 11-10 258,649.

- Tranco is a 30-day rolling average. Real traffic began rising around **mid-September** and accelerated through October. It has grown more slowly since November (11-03 287,037 → 11-10 258,649).

**CrUX** (traffic database, crux_domain, bucket levels only):

- India has been in the top 1M since **2026-06** , which means real Chrome users within about 4 months of launch.

- The US, Canada, Australia and the Philippines entered the top 500k in August; the UK went from top 1M in September to top 500k; the Netherlands entered the top 500k in September; Poland and Vietnam entered the top 500k in October.

- The global bucket first reached the top 1M in October.

- Coverage is all-English and near-English markets, consistent with a traffic mix driven by English-language Google long-tail search.

**On-site counts** (from its public API `/api/v1/posts`, 2026-11-11 pulled all 1,515 questions):

- Since the 10-29 conversion, about **103k views** in total, averaging roughly 8k a day over 13 days. That matches the few-thousand-visits-a-day scale you'd expect at Tranco rank 250k.

- **Pages migrated on 10-29 took 71,947 views from 181 pages (about 70%)** . The top 20 alone account for 47k views: Walmart Spark driver earnings 8.3k, Facebook Protect 4.4k, deleting a Meta account 4.1k, DoorDash earnings 4.0k, Android virus pop-up 3.6k, localhost connection refused 3.3k, whether Instagram charges 3.2k…

- Since 11-04, 150–220 new items a day, and each new item gets only single or double-digit views.

**Comparable sites over the same period** (Tranco, 08-06 → 10-01 → 11-10): Quora 474→505, answers.com 4,210→4,613, ask.com 5,778→7,565, reference.com 5,578→6,185, wikiHow 1,818→2,077, Brainly 23,886→19,641, all flat or declining. AI FAQ sites in the IAC/System1 family are falling sharply: faqtoids.com 140,629→472,341, questionsanswered.net 267,444→546,903. The category as a whole is being squeezed by AI Overviews. unanswered.io's rise is its own, not a tailwind from the sector.

**App** : no app with the same name in the App Store. All of its traffic is on the web.

### II. How it became what it is now (Wayback snapshots + RDAP + site API)

| Timeline | Status | Evidence |
|---|---|---|
| 2026-01-31 | Domain registration | RDAP (identitydigital) registration 2026-01-31; no Wayback snapshots before 2025 years ago, so it looks like a new domain |
| 2026-03-04 | "The Internet's Question Catalog": **75,711+ guides, 54 categories** , claiming expert review and an average rating of 4.7 stars | Wayback homepage snapshot |
| 2026-04-08 → 08-28 | Guide count stalled at **81,088** , adding about 5,400 articles in a month, then stopped growing | Wayback 04-08 / 05-06 / 07-26 / 08-28 snapshots |
| 2026-09-29 | Rebranded as "Independent, source-backed explainers", with author pages showing licensed-professional personas (e.g. "Julian Moreau, LMFT") | Wayback 09-29 snapshot; those author pages are all 410 now |
| 2026-09-24 → 10-08 | Google September 2026 spam update | Google Search Status Dashboard |
| 2026-10-28 / 10-29 | Domain record changed and DNS moved to Cloudflare; the new Q&A system launched on 10-29, with 181 questions bulk-loaded the same day | RDAP last changed 10-28; the API has 10-29 with 181 entries, all authored by `unanswered` (id 4,10-29 20:32 registered) |
| From ¥10-29 | Spot check of 80k `/guide/` pages: ones with outside links and traffic 301 to the new Q&A pages, the rest return 410 Gone | I checked 26 old URLs: 4 301 (all previously linked from Facebook groups or outside sites), 22 410 |
| From ¥11-04 | 150–220 new Q&A items a day, bylined to 13 "staff writer" accounts | The API breaks down by date; the sitemap holds 1,095 question URLs in total, all with lastmod after 10-29 |
| 2026-11-10 | Google's November 2026 core update began rolling out | Google Search Status Dashboard |

**The publishing pipeline is public.** Its OpenAPI (`/api/openapi.json`) includes `POST /api/v1/publish`: "Publish an editorial question and answer", authenticated with a bearer token and idempotent per key. Fields include `writer` (the staff writer username), `sources`, `verified`, `review_after`, and `legacy_paths`: "Old /guide/ URLs this page replaces; they redirect to it while it's in search". In other words, the old-guide migration and bulk publishing of new answers were both done by scripts calling this endpoint.

**"Staff writer" is a shared byline.** The editorial standards page (updated 2026-10-31) states: "Staff writer is a house byline used by the Unanswered editorial team … we use AI tools to help with research and drafting". 13 bylines (James Whitfield, Maya Okafor and others) each carry 40–120 answers, all with a reputation of 1. The only real human admins are `rehmat` (Rehmat Alam) and `qasimkhan` (joined 10-31).

**About the author** : Rehmat Alam (rehmat.works) built his name on large-scale scraping. He founded ProAPIs, Inc. in 2020, which runs the Google search results API **SerpsBot** and iScraper, a professional social data API. iScraper was shut down after 2025-10 sued it over scraping, and settled in 2026-09-16. In 2025 he founded Zyref, LLC, which now focuses on Mezaic (a personal profile / publishing platform) and GovernKit. **Neither his personal site nor the Zyref site lists unanswered.io** , so this is a low-profile project. His SERP scraping background explains how the site sources its topics: "We also start from what readers search for most … labelled Popular question: they come from search demand".

### III. Competitors

| Product | Form | Launched | Scale (Tranco 11-10) | How unanswered.io differs |
|---|---|---|---|---|
| Quora | Web / App | 2009 | 505 | Many people answer, UGC-driven, with uneven quality. unanswered is one answer per question, editor-bylined, with sources and a review date |
| Reddit (including Reddit Answers) | Web / app / big-company features | 2005 / 2024 | 109 | Discussion threads, which Google has ranked heavily over the last two years; unanswered's pitch is "no digging through threads" |
| answers.com / ask.com / reference.com | Web | Long-established | 4.6k / 7.6k / 6.2k | Established Q&A / content sites, large but not growing; unanswered is a low-cost AI-era copy |
| wikiHow | Web | 2005 | 2,077 | Leans toward how-to guides with real-person testing. unanswered covers a broader range (earnings, bills, platform policy) |
| Stack Exchange | Web | 2009 | 1,086 | Professional community. unanswered's reputation / accepted-answer / bounty mechanics clearly copy it (the API has bounty and noaccepted sorting) |
| Brainly / Gauth | Web / App | 2009 / 2020 | 19,641 / 15,361 | Homework Q&A aimed at students |
| faqtoids.com / questionsanswered.net / consumersearch.com | Web | Old guard (System1 / IAC family) | 470k / 550k / 130k | Also batch-generates FAQs from search demand, monetized with ads, and is being demoted by Google. unanswered is an ad-free "new shell" |
| answerbag / blurtit / askmefast / theanswerbank | Web | 2000s | 130k–700k | Old Q&A site, basically stalled |
| Google AI Overviews / ChatGPT / Perplexity | Big-company features | 2023–2024 | — | Gives answers directly in search results, the biggest threat to this kind of site. unanswered's "sources + review date + one answer per question" format is built to be cited by them |

This is a crowded category that is shrinking overall. unanswered.io is much smaller than these established sites. What sets it apart is that it's new, and it's rising against the trend.

### 4. Where users come from

1. **Google organic search (main channel, strong evidence, but no direct referral data obtained)**
    - Topics are all English long-tail search terms: earnings, bills, platform policy, device troubleshooting, and clearly in high-CPC territory.
    - CrUX distribution is concentrated in English-speaking markets: US, UK, Australia, Canada, India and the Philippines.
    - 70% of on-site views land on pages migrated from old guides 301.
    - The site itself says it starts from search demand, and that only editor-approved pages go into the sitemap.
    - No ads and no social accounts. 0 mentions on X (x-radar, 120 days), which rules out the other main channels.

2. **AI answer citations (weak evidence)**
    - A few external traces: the Instagram reel caption has "unanswered.io unanswered.io +1", the Apple podcast description and the squier-talk forum post both contain "unanswered.io", which looks like users copied it along with the citation chip from an AI answer.
    - But seomonitor's AI visibility report (US region) shows "No visibility found". The recall card says "LLM citation index rank #869", but I couldn't find the original page, **unverified** .

3. **Sporadic external citations (small)** : a Facebook group post, and a medm.com press release ("Full story: unanswered.io/guide/best-health-tracking-apps", a guide that is now 410). Guides that got cited mostly survived and were 301.

4. **Community / retention features (just launched, almost no traffic)** : the daily quiz leaderboard says "No scores yet"; user IDs go up to 510, and many of the roughly 500 accounts are link-spam accounts (webtechs, techiehubs, Asmorix_Technologies); 420 questions have no answers. The email quiz reminder goes out daily at 12:00 UTC, but very few people are likely subscribed.

**Not found** : paid ads (the page only has GA4 G-32EYGL54Q8, no AdSense or other ad code; our Meta / Google / TikTok ad libraries don't index it either), Product Hunt / HN, influencers, press coverage.

### V. Why it took off

1. **Mass-produced ~80k long-tail guides with AI and caught Google's long-tail traffic.** Over 2 months it scaled to ~81k pages; India in June, and the US, UK, Australia and Canada entered CrUX in August. Traffic clearly rose in mid-September, pushing Tranco from off the list to 258k.

2. **Swapped the shell before Google cleaned up bulk content.** After the September spam update, it cut the fake-credential author persona and ~80k thin pages (410), kept about 180 winners, migrated them with `legacy_paths` 301 to concentrate link equity, and switched to an E-E-A-T-friendly format: one question, one answer, bylines, sources, review dates, revision history, a public editorial policy and an AI-use disclosure. Now 70% of views still come from these migrated pages.

3. **The author builds a SERP scraping API (SerpsBot).** He can get "what people are searching" cheaply and has the ability to build a fully automated publishing pipeline, posting 200 entries a day.

4. **The topics are questions AI Overviews haven't fully absorbed.** Time-sensitive platform policies, earnings, bills and device faults (Walmart Spark pay, Facebook Protect, Sam's Club membership) need facts "as of a date", where answers with review dates have an edge.

### VI. Differentiation

Compared with Quora and Reddit: one question, one answer, editor-owned, with sources and review dates, no digging through threads. Compared with System1-style FAQ sites: no ads, no affiliate links, and a published editorial policy. Compared with AI Overviews: it's a "citable source" for AI and search engines.

But **these differentiators are not why it grew** . The traffic came from the AI-generated guide mass that piled up before the redesign; the new format is there to **hold onto** and **legitimize** that traffic, and the Q&A community and daily quiz haven't brought visible volume yet. The real "edge" is in the operations: mass-produce → filter winners by traffic → swap to a Q&A shell and migrate link equity in one go, all automated.

### 7. How many people have seen it

Few people know about it yet:

- 0 mentions on X (x-radar, 120 days);

- no press coverage, no Product Hunt or HN;

- the author's own homepage and company site don't list this site;

- the only findable results are automated safety-score pages like scamadviser and gridinsoft, and seomonitor's empty report;

- revenue isn't public, and monetization hasn't started.

**Copycats** : none found copying the "AI mass guides → Q&A shell" playbook. Programmatic FAQ sites are an old, crowded category in themselves.

### 8. What's worth borrowing

- **Mass-produce first, pick winners by real traffic, then concentrate and rebuild.** Of the ~80k pages, keep about 180, and 410 the rest, to concentrate link equity; the publishing API gets a `legacy_paths` field so migrations and redirects run automatically.

- **Format for AI and search engines** : a direct answer up top + detail + a source list + "checked / review by" dates + a public revision history + an editorial policy (explicitly disclosing the house byline and AI assistance). This combination is safer than inventing credentialed personas.

- **Only put editor-approved pages in the sitemap and allow indexing** ; noindex the rest to control the site-wide quality score.

- **Pick time-sensitive topics like platform policies, gig income and bills** , and review and refresh regularly to capture "as of today" demand.

### 9. What I'm still unsure about

- Why traffic started rising in mid-September: did rankings climb after the August spam update, was it the September redesign, or was it AI citations? We don't have Search Console or referrer data, so we can't tell. How much each of Tranco's five data sources contributes is also unclear.

- **The Google November 11-10 core update, starting 2026, could wipe it out** . New rebuilt pages get only a few to a few dozen views each, and there's no evidence yet that they can pick up the old guides' traffic.

- The monetization path is unclear: no ads or affiliate links right now. It might add ads later, be building a "citable asset", or funnel users to other Zyref products, **unverified** .

- "LLM citation index #869" couldn't be traced to an original source; seomonitor's US report shows no AI visibility.

- How many of the on-site view counts come from crawlers is unknown.

### Sources

- unanswered.io homepage, /about, /about/how-it-works, /about/editorial, /writers, /quiz, /write, /users, /api/openapi.json, /api/v1/posts (full crawl, 2026-11-11), /sitemaps/questions.xml

- Wayback Machine homepage snapshots: 20260304, 20260408, 20260506, 20260726, 20260828, 20260929, 20261031

- RDAP: rdap.identitydigital.services (registration 2026-01-31, last changed 2026-10-28); crt.sh (earliest certificate 2026-02-10)

- traffic library rank_day / crux_domain (as of 2026-11-10)

- Google Search Status Dashboard ranking update history; Search Engine Journal "Google September 2026 Spam Update"

- rehmat.works (About / Work / Updates / Blog "iScrape chapter closed officially"), zyref.com

- Jina search "unanswered.io": scamadviser, gridinsoft, seomonitor, Facebook group post, Instagram reel, medm.com press release

- x-radar search "unanswered.io" (120 days, 0 posts)

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.
