This page records the creator’s observations and decisions while building SarvKit, including approaches that were later reconsidered. It is a dated account, not a controlled experiment or an independently verified diagnosis of Google’s ranking systems.
Two things you should know before reading. First, I am deliberately not publishing absolute traffic figures, per-page keyword data, or revenue. What is here are ratios, rates and the reasoning — enough to be useful, not enough to be a competitor's roadmap. Second, this page is updated as the site ages rather than rewritten, so entries stay as they were written even when a later entry proves them wrong.
Current reading note: The dated entries below retain the interpretations recorded at the time. Domain age alone does not make a site invisible, and page count or suite breadth alone does not establish doorway or scaled-content abuse. Link attributes do not retroactively cancel months of work. Observed citations, impressions and word-count buckets do not prove visits, completed actions or a causal ranking rule. Treat private analytics as attributed observations, not independently verified evidence.
The timeline
Breadth-first, on purpose
The opening decision was to build wide rather than deep: thirteen unrelated suites covering finance, health, legal, developer, design, PDF and study tools. The reasoning was that I had no idea which category would work, and static client-side tools cost essentially nothing to keep online — no servers, no per-tool maintenance.
That reasoning still holds, with one correction I would not understand for another four months. Breadth is a fine way to find a winner. It is a bad way to keep spending once you have found one.
Roughly 500 pages published in a few weeks
Nearly the whole site went live in a short window. Every suite received a comparable number of guides — about twenty each — regardless of whether that suite had shown any sign of working, because at that point none of them had any data at all.
This is the single most expensive mistake in the log. The tools were cheap. The guides were not: they were the bulk of the actual work, and they were spread evenly across thirteen categories before a single piece of evidence existed about which category deserved them.
Bing arrived first. Google did not arrive at all.
Within weeks Bing was indexing pages and sending real clicks. Google, given the identical sitemap, the identical markup and the identical content, sent essentially nothing.
Nothing was wrong technically — no blocked resources, no canonical errors, no robots directives. The engines simply have very different tolerance for an unknown domain. Bing will index a new site liberally and let ranking sort it out. Google largely declines to spend crawl budget until it has a reason to.
AI citations started growing before clicks did
Bing Webmaster Tools reports how often an AI assistant cites your pages when composing an answer. That number began climbing steadily, across a widening set of distinct pages, well before those pages earned meaningful click traffic.
This is a genuinely new measurement problem. A page can be doing real work — being read, being used as a source, shaping an answer someone acted on — while its click count sits at zero. Judged on clicks alone, those pages look like failures.
The first per-page data: brutal concentration
The first proper per-URL export was the most informative thing that happened all year. Traffic was not spread thinly across 200 tools. The overwhelming majority of all clicks landed on one tool, and it was not one I would have predicted — nor one I had spent much time on.
The winner was a sample file generator: you specify a format and an exact size, and it hands you a valid file. Working out why that one won, when 200 others didn't, became the central question.
The answer turned out to be simple. It produces an artifact. An AI assistant can tell you an email attachment limit instantly and correctly — which is why our guide explaining that limit earns nothing. It cannot hand you a 25 MB file to test the limit with. The click survives only where the answer is a file rather than a fact.
Rejecting the obvious growth plan
The tempting next move was to build a landing page for every popular variant — one for each size, one for each format, a dozen or so near-identical pages. Search demand clearly existed for each.
I didn't, and I think that was right. Those pages would have been about ninety percent identical to a page that already ranked for the same queries. That is the textbook doorway pattern, and it is exactly the profile that gets a large templated site into trouble.
The alternative was depth: make the one winning page do more. Multi-page documents, more formats, real image dimensions, larger files, exact byte sizing. Same demand captured, one URL, no duplication risk.
Testing with strict parsers found bugs nothing else caught
Generating a file of an exact byte size that is also valid is much harder than it looks, and the failures are silent. Padding to hit a target size is where everything breaks. Real bugs found only by validating output with strict library parsers rather than by opening files in a viewer:
- PDF: a content stream declared a hardcoded
/Lengththat did not match its actual byte count. Lenient viewers rendered it fine; strict readers over-read past the stream and rejected the file. Byte offsets in the cross-reference table have to be computed, never assumed. - DOCX and XLSX: padding bytes were appended after the closing XML tag. Word opened them; proper XML parsers refused them outright with "extra content at end of document" — which broke the exact audience that needs generated test files. The fix was to place padding inside an XML comment.
- PNG: pad using a
tEXtchunk inserted beforeIEND, so the chunk structure stays valid. - JPEG: pad using comment markers placed after the start-of-image marker; padding near the end decodes unreliably.
- Structured text: JSON, XML and SVG each have a minimum size below which no valid document exists. Requests below that floor were silently returning whitespace.
Google finally showed up — as impressions, not clicks
Google impressions went from essentially nothing to a genuine, sustained presence. Clicks stayed near zero, because average positions were well outside the pages anyone actually looks at.
It is tempting to read that as failure. It isn't. Impressions arriving before clicks is the normal sequence: it means the pages are indexed and being considered. What was interesting is where it happened — on a handful of country-specific financial calculators, not on the tool that dominates elsewhere.
The mistake: noindexing almost the entire site
Acting on widely-repeated forum advice that search engines were penalising sites with large volumes of AI-assisted pages, I applied a noindex header to nearly every page, keeping only a couple of proven ones indexed. The theory was that a smaller, higher-quality indexed footprint would lift the rest.
It was reverted within days. Four reasons it was wrong:
- Most non-indexed pages had never been crawled in the first place. Removing something that was already absent changes nothing.
- It de-indexed the pages that were indexed — including the ones that had just started earning impressions.
- There is no documented mechanism by which fewer indexed pages improves how a site is assessed. It is folk wisdom.
- The header included
nofollow, which silently cancelled months of deliberate internal linking, and the directive applied equally to the engine that was actually sending traffic.
Comparing the same weekdays before and after showed no measurable damage, which was luck rather than judgement.
What actually predicts an AI citation (it isn't length)
With citation data for every guide on the site, I could finally test the assumption everyone repeats — that longer, more comprehensive content wins. I bucketed all ~240 articles by word count and measured what share of each bucket had ever been cited:
| Article length | Share of articles ever cited |
|---|---|
| Under 600 words | 46% |
| 600–799 words | 44% |
| 800–1,199 words | 50% |
| 1,200–1,799 words | 51% |
| 1,800+ words | 33% |
Flat — and the longest bucket performs worst. Two of the most-cited pages on the entire site are among the shortest things published, at just over 500 words each.
The real pattern showed up when comparing suites with identical median article length. One suite of unit-conversion references (median 637 words) and one suite of fitness guides (median 649 words) differ by roughly two orders of magnitude in citations. Same length, same site, same author, same template.
The difference is what the content is. Bounded reference material — viscosity units, paper size specifications, PDF format differences, refund-policy rules — gets cited constantly. General advice — intermittent fasting, resume tailoring, tone of voice — gets cited almost never, despite far higher search demand.
Another Google visibility drop: investigation before retirement
Google impressions collapsed again after a brief increase, and clicks remained near zero. Search Console showed no reported manual actions or security issues. The inspected mortgage calculator URL had a successful Googlebot smartphone fetch on 8 October, with crawling and indexing allowed and matching declared and Google-selected canonicals. Its sitemap discovery section showed a temporary processing error. Those observations did not establish the cause of the traffic loss or prove that the entire site was technically healthy.
I suspected that expertise-sensitive tools and guides were holding the site back. Reviewing Google's guidance did not confirm an E-E-A-T penalty, and reaching month seven does not mean a documented six-month sandbox has expired. Removing pages can still sacrifice backlinks, useful content and future recovery opportunities even when current clicks are negligible.
Five suites removed, with a recoverable backup
The retired suites were CalcWise, FitCalc, InsureKit, LegalKit and UnitSnap. Their tools, guides and source directories were removed, along with their deployment-output copies. A recoverable backup was kept outside the site; historical analytics and marketing drafts were preserved and remain excluded from deployment.
- Inventory: 132 registered tools remain across eight suites: DevKit (37), PDFTools (16), ResumeKit (5), TextKit (17), SEOKit (9), DesignKit (24), BizKit (14) and StudentKit (10).
- Discovery and navigation: removed retired registry entries, sitemap URLs, search entries, manifest shortcuts, shared navigation, no-script navigation and public recommendations. Removed the developer-suite Byte Converter entry that pointed into a retired suite, and changed Receipt Scanner recommendations from GST to the retained Sales Tax Calculator.
- Site content and tooling: updated homepage, about page, READMEs, contact and policy scope, 404 recommendations, branding and active strategy documentation. Removed retired-only scripts and tests, and regenerated search, navigation, branding and deployment output.
- Cache refresh: advanced the service-worker cache to
sk-v13and the social-preview image version to4. - Coupled fixes: cleaned up duplicated malformed no-script navigation and strengthened encoding checks with strict UTF-8 decoding, retained-icon assertions and exact HTTP/source byte comparisons.
The rebuilt search index contains 335 entries: 160 tool-category pages, 167 guides and eight suite roots. Indexed tool-category pages include resume templates and examples, so that count is not the same as the registered-tool count.
Checked the removal beyond the visible navigation
The following checks passed after retirement and before this timeline update. They validate the local source and deployment output, not Google's indexing decisions or the live deployment.
- Regression tests: 71 main-suite tests passed; targeted retirement, offline and palette checks passed eight runner tests without skips, with 14 palette assertions. Earlier navigation, discovery and service-worker checks passed 32 tests.
- Browser navigation: 80 cases passed across 20 pages, desktop and mobile, with JavaScript enabled and disabled.
- HTTP retirement: all eight retained suite roots returned 200 locally; 25 retired-route variants returned real 404 responses without redirects.
- HTML links: source and deployment scans each checked 341 pages and 9,503 internal links, finding no broken, orphaned or unreachable pages. Every checked page was reachable within two clicks of home.
- README follow-up: found and removed 20 missed retired-suite links across four suite READMEs, corrected three outdated inventory totals and fixed a malformed root README link. The follow-up scan checked 12 active Markdown files and 450 local links with no remaining issues.
- Discovery files: source and deployment sitemaps each contained 340 matching, unique URLs with no missing destinations or retired-suite URLs. Robots sitemap declarations, manifest targets and all 335 search-index destinations resolved locally in both trees.
24 October 2026 — wording review and local checks
Reviewed all 341 remaining public HTML pages against their tool implementations and Google's helpful-content, title-link, snippet and spam guidance. Wording changed on 330 pages: 5 shared pages, 97 DevKit/TextKit pages, 113 BizKit/StudentKit/ResumeKit pages and 115 PDFTools/DesignKit/SEOKit pages. The ownership-verification file is not a public content page and was excluded.
- Removed or qualified unsupported expertise, outcome, performance and privacy claims; aligned titles, descriptions, FAQs and instructions with actual features and limitations.
- Clarified project attribution without unsupported employment or experience claims, and explained email handling without guaranteed response times. Public attribution now uses the SarvKit brand without personal names or social-profile links.
- Clarified illustrative ROI assumptions, heuristic email-subject feedback, PDF compression trade-offs and local SEO checks. NATO output now uses the official “Alfa” spelling while still accepting legacy “Alpha” input.
- Corrected contradictory behavior and added regression coverage. Added Organize PDF to the PDFTools homepage's main cards and corrected its count to 16 tools.
The combined wording and behavior regression run passed 103 runner tests. These are selected checks, not certification of every input, browser or third-party workflow. Historical entries remain records of their time; publication dates, private metrics and earlier testing claims still need owner evidence where not independently verified.
Final local integration checks passed: 25 affected wording/navigation tests, 71 template/include/build/HTTP checks, and 80 desktop/mobile browser-navigation checks each against source and generated output. Both link scans covered 341 pages and 9,504 internal links with no broken links, unreachable pages or orphan pages. Regenerated the search index and deployment output, verified the build, and aligned sitemap modification dates with the 330 changed pages.
This work is local and has not been deployed. Wording improvements do not establish the cause of the traffic loss, constitute Google approval or guarantee ranking recovery.
What I would tell someone starting today
- Build cheap and wide only until something moves. Then stop and concentrate. The expensive mistake is not breadth — it is continuing to invest evenly after the data has spoken.
- Build for a useful outcome. Downloadable artifacts can give people a reason to visit, but an AI-answerable query does not make clicks impossible, and artifact tools do not guarantee demand.
- Do not judge pages by clicks alone. Citations and impressions are real signals, and on a young site they arrive first.
- Choose the format that serves the task. Clear reference material and well-supported advice can both be useful; this site’s observations do not establish a universal preference for one format.
- Validate important outputs with suitable tools and human inspection. Visual appearance alone does not verify structure, accessibility or correctness; this does not mean every apparently correct file is broken.
- Be slow to restructure. The earlier broad noindex experiment was reversed. October's suite retirement is a separate scope decision whose search outcome remains unmeasured, not proof that pruning works.
- Do not treat site age as a deadline. In the seventh month, Google visibility remains volatile. This log does not establish a fixed sandbox period or a guaranteed recovery date.
Still open
Questions this log has not answered yet: what caused the October visibility collapse; whether retiring five suites changes discovery or search performance after deployment and recrawling; whether Google impressions return and convert into clicks; whether reference-style content can be produced deliberately; and whether a second artifact tool can work. Deployment, live 404 and sitemap checks, Search Console follow-up and post-change measurement are still pending. A later recovery would not by itself prove an E-E-A-T diagnosis. I will add dated entries as these questions resolve — including if they resolve badly.