Case study · kwestiaprompta.pl

A brand-new site: 0 → 98.6% indexed in 5 months

kwestiaprompta.pl launched in March 2026 with 336 published articles and nothing in Google's index. The only indexing intervention ever used: IndexFixer and its dynamic internal-linking widget. It's our own property — which is exactly why we can show the full, unedited history from day zero.

442

URLs monitored

436

indexed today (98.6%)

296

URLs surfaced by the widget

98%

of widget URLs now indexed

The five-month picture

Column definitions: Indexed and Not indexed are Google's verdicts from the URL Inspection API. The remainder (Other) is pages Google reports as discovered/excluded plus pages not yet checked in the current cycle — which is why the three columns don't sum to the total.

View the data as a table (weekly)

Date	Total URLs	Indexed	Not indexed	Other*
2026-03-12	336	0	0	336
2026-03-19	395	1	0	394
2026-03-26	396	1	0	395
2026-04-02	396	40	1	355
2026-04-09	396	183	15	198
2026-04-16	396	198	16	182
2026-04-23	405	227	17	161
2026-04-30	421	263	16	142
2026-05-07	431	292	15	124
2026-05-14	431	292	15	124
2026-05-21	431	310	10	111
2026-05-28	431	332	10	89
2026-06-04	442	344	9	89
2026-06-11	442	355	9	78
2026-06-18	442	383	8	51
2026-06-25	443	411	24	8
2026-07-02	443	413	22	8
2026-07-09	443	391	44	8
2026-07-16	443	411	25	7
2026-07-23	443	414	23	6
2026-07-30	443	414	22	7
2026-08-06	442	437	1	4
2026-08-13	442	436	1	5
* discovered / excluded / not yet checked in the current cycle

Three honest details worth noticing. First, the early weeks are slow — a fresh domain has no reputation, and no tool changes that. Second, the July wobble: 17 pages dropped out of the index mid-July and every one of them was pulled back within two weeks — dropped pages re-enter the widget automatically, get re-discovered, and return. Indexation is not a one-time achievement; it's a state you hold. Third, the sharp early-August jump (414 → 437 indexed in about a week) is partly a measurement artifact: on August 1 the site was migrated to IndexFixer's new server-side inspection engine, and the full re-scan that followed cleared a backlog of stale statuses. Google didn't index 21 pages overnight — some of them had been indexed for a while and the old data hadn't caught up.

The setup

Site: kwestiaprompta.pl — Polish content site about AI tools and prompting (new domain, no backlink profile, no ads, no manual URL submissions).
March 12, 2026: IndexFixer installed; the "not-indexed posts" widget added the same evening, sitewide under every article.
That's it. No Search Console manual requests, no indexing services. The widget surfaced up to a handful of not-indexed URLs at a time as plain HTML links, ordered by neglect (never-crawled first), swapping automatically as pages got indexed.

What the widget actually did

Over five months, 296 of the 442 URLs (67%) cycled through the widget — Google simply hadn't got to two thirds of this site yet. Of those 296, 290 are in the index today (98%). Some concrete trajectories:

An article published at launch entered the widget on day one and sat unindexed for weeks while the domain earned trust — then indexed. The loop doesn't promise speed on a cold domain; it promises Google keeps finding the page until it sticks.
/chatgpt-dla-uczniow-jak-zaczac-krok-po-kroku-poradnik/

A post entered the widget on June 21 and was indexed three days later.
/claude-opus-vs-sonnet-vs-haiku-ktore-ai-wybrac-porownanie-modeli/

A post entered the widget on July 13 and was indexed two days later.
/elevenlabs-jak-uzywac-ai-do-klonowania-glosu-i-text-to-speech/

Is this just "new sites index anyway"?

Partly — which is why we also measured the widget against a comparison group across our whole network of 12 production sites (same sites, same time window):

Metric (same clock for both groups: from publication date)	URLs surfaced by the widget	Comparable not-indexed URLs, no widget
Not-indexed → indexed	68% (2,278 / 3,329)	41% (3,350 / 8,257)
Median days from publication to indexed	35	38
Indexed within 30 days of publication	42%	38%

The conversion gap is the headline: 68% vs 41%. The time gap measured from publication is modest — most of a neglected URL's wait happens before anything can act on it. The speed story lives elsewhere: once a URL enters the widget, half are indexed within 3 days (median; 59% within a week) — and these are pages that had already been waiting for weeks. We report both clocks so you don't have to guess which one flatters us.

And the selection bias runs against the widget: it deliberately receives the most neglected URLs — pages Googlebot never crawled or abandoned longest — i.e. the hardest cases.

Methodology & honest limits

This is an observational comparison on production data, not a randomized experiment. Correlation isn't causation; we report what happened, we don't promise outcomes.
"Transitioned" in the comparison group = status changed to indexed more than a day after first being seen — an approximation that may include edge cases (e.g. stale-404 recoveries).
Clocks: the comparison table measures both groups from the publication date (the only start both share). The "3 days" figure uses a different clock — time from widget entry — and is deliberately not compared against the control group.
This case is most representative of content sites publishing at volume on a young or under-linked domain — where the discovery problem is extreme. A 50-page service site has a different problem profile and shouldn't expect these curves.
Google indexes what it judges worth indexing. The widget removes the discovery bottleneck; it cannot fix thin or duplicate content. How the loop works — and what it can't do.
Data through August 13, 2026, from IndexFixer's own status history (Google Search Console URL Inspection API). We'll rerun the numbers periodically.

Run the same loop on your site

Two-click Search Console connection. No API keys, no Google Cloud. €50/year, 14-day refund.

Get IndexFixer Pro