Skip to content

brogav.com — estimating traffic with no paid tools

brogav.com is a micro-traffic B2B site — roughly 1,200 monthly visits — that became the validation target for a traffic-estimation method built entirely from free, public data sources. No SimilarWeb Pro, no Ahrefs, no SEMrush. The output is a quantitative traffic and keyword profile with explicit confidence ratings, produced by triangulating five independent sources (June 2026).

What this proves

  • You can estimate any site's traffic for free — including competitors and prospects — if you triangulate rather than trust one source.
  • Confidence is part of the answer. A single estimate is a guess; three independent sources agreeing is a finding. The method makes you state which one you have.
  • Precision degrades predictably at the bottom. Below roughly 5,000 monthly visits, the free panel-based estimates get unreliable in knowable ways — and knowing that is what keeps you from over-claiming.
  • Competitor keyword data is the real prize. The most valuable output usually isn't the target's visit count; it's the search volumes their competitors rank for.

The five sources

Each answers a different question, and none is trustworthy alone.

Source What it gives you Where it fails
SimilarWeb free API Monthly visits (3 months), global/country rank, bounce rate, pages/visit, traffic-source split, top keywords with volumes and CPC, top countries Unreliable below ~5k visits/mo; may return 403/504 or empty for small sites; throttles after ~5–10 queries
Wayback Machine CDX API Every archived capture with timestamp and status — a proxy for crawl frequency, site age, and page-count growth Crawl frequency reflects archivist interest, not only popularity; rate-limited to roughly 15 req/min
Tranco list Whether the domain ranks in an aggregated top-1M list Binary at this scale — absence just means "outside the top million", which is most sites
Cloudflare Radar DNS-based popularity bucket (top 100 / 1K / 10K / 100K / 1M) 403 means "not ranked"; automated access can be challenged
Google Trends + web search Relative interest over time; third-party estimates, PR mentions, industry reports Trends is browser-only (no free programmatic API) and relative, never absolute

The method

1 · Pull the panel estimate. Query the SimilarWeb free endpoint for the target. Extract visits, rank, engagement, traffic-source split, and top keywords.

curl -s "https://data.similarweb.com/api/v1/data?domain=example.com"

2 · Establish age and trajectory from the archive. The CDX API gives first-capture date (when the site became visible to crawlers) and monthly capture volume (a rough activity/complexity trend).

curl -s "https://web.archive.org/cdx/search/cdx?url=example.com/*&output=json\
&fl=timestamp,original,statuscode,mimetype&collapse=timestamp:6&limit=500"

Count captures per month, count unique URLs, and read the shape: rising unique-URL counts mean the site is growing; a flat line for years means it isn't.

3 · Bound the magnitude. Tranco and Cloudflare Radar don't give you a number — they give you a ceiling. A domain absent from both is confidently outside the top million, which for a micro-traffic site is the expected, corroborating result.

4 · Mine competitors for keyword intelligence. This is the step people skip and shouldn't. Query the same free endpoint for three or four competitors: their keyword lists reveal what buyers in the category actually search for, with real monthly volumes and CPCs — intelligence the target site alone can never give you.

5 · Triangulate and rate confidence.

Confidence Criterion
High 3+ independent sources agree on the order of magnitude
Medium 2 sources agree, or 1 strong source with corroborating context
Low A single source, or sources that disagree — report as a range, and say so

What it looked like for brogav.com

A micro-traffic B2B site at roughly 1,200 monthly visits: present in panel data but near its reliability floor, absent from both Tranco and Cloudflare Radar (consistent, and expected at that size), with an archive history establishing site age and a modest page count. The individual signals were each weak; agreeing on the same order of magnitude is what made the conclusion usable.

The strategic read matters more than the number. At ~1,200 visits/month you are not going to win contested head terms — the keyword strategy chapter's vacant-niche argument applies directly. Micro-traffic sites win with specific, low-competition, high-intent content, and the traffic estimate's real job is to set that expectation honestly before anyone budgets for a head-term campaign.

The numbers

  • ~1,200 monthly visits (the validated target figure, June 2026)
  • 5 independent free sources triangulated
  • ~5,000 visits/month — the approximate floor below which panel-based estimates become unreliable
  • 3+ agreeing sources required for a high-confidence rating
  • Unmeasured: the true visit count. There was no analytics access to validate against — which is precisely the situation the method exists for, and precisely why confidence tiers are mandatory rather than decorative.

What worked

  • Triangulation over authority. No single free source is trustworthy at this scale; agreement across independent methodologies is.
  • Stating confidence explicitly. It converts an estimate into something a reader can act on proportionally.
  • Competitor keyword mining. Consistently the highest-value output, and free.
  • Negative results as evidence. "Absent from Tranco" is a real, corroborating data point, not a failed query.

What failed

  • Trusting traffic-source splits at micro scale. Below the reliability floor, channel attribution can read 0% across the board — an artifact, not a finding.
  • Expecting precision. These are order-of-magnitude estimates. Reporting "1,247 visits" from panel data would be false precision.
  • Rate limits. Both the panel API and the archive throttle; batch queries need pacing.

What we'd do differently

  • Pull competitor keyword data first — it's the most valuable output and doesn't depend on the target resolving.
  • Record the raw JSON responses with their date, since these free endpoints change shape and rate limits without notice.
  • For any site you actually control, none of this is necessary: Search Console gives you ground truth. This method is for sites you don't own.

Chapters this case feeds

Lesson Chapter
Triangulate free sources; rate confidence explicitly Measurement and baselines
Which free tool answers which question Tool directory
Micro-traffic sites win niches, not head terms Keyword and SERP strategy
Own the site? Use ground truth instead Google Search Console