Skip to content

Methodology

How every number is calculated

Each estimate is a range built from signals we actually observed, with the inputs shown next to it. When a signal is missing, that part of the estimate is left out and the confidence drops — nothing is filled in.

In the databaseCount
Funnels tracked7,445
With ads running now6,245
Ads tracked93,643
Ads with EU/UK reach disclosed by Meta7,163

Live counts, data as of November 19, 2026.

Where the funnels come from

Every day we search Meta’s public Ad Library for buying-intent phrases in each niche, open the page each ad links to, and record the advertiser, their ads (start dates, versions, platforms, copy) and the page itself (type, offer, price, builder, tracking pixels). Known advertisers are re-checked daily, so a funnel’s history — ads added, ads stopped, price changes — is real, day by day.

Traffic

In the EU and UK, Meta discloses how many people each ad has reached. That is our strongest signal. People reached per day (reach ÷ days running) becomes monthly visits through three named bands:

StepBandBasis
Impressions per person1.3–2.5impressions per person reached over a campaign (Meta delivery norms: 1.3–2.5)
Link click-through rate0.6%–1.6%link clicks per impression (Meta feed ads benchmark: 0.6–1.6%)
Clicks that load the page70%–90%link clicks that finish loading the page (70–90%)

The central value uses the middle of each band (about 1.40 visits per 100 people reached) and the range shown is that value halved and doubled. An ad shown only in the EU/UK is fully observed. An ad also shown elsewhere gives a floor: the rest of its delivery isn’t disclosed, and we say so.

US-only ads disclose no reach. They get a modelled figure only once at least 100 fully disclosed ads exist to model from — until then they have none. Domain popularity (Tranco rank, and Cloudflare Radar or Chrome UX Report when available) caps or supports the result; search traffic is shown as its own channel when measured.

Ad spend uses the same reach with public CPM bands: EU $4–$10, UK $6–$14, US $10–$25 per 1,000 impressions.

Revenue

Monthly revenue = visits × the share of visitors who buy (by funnel type) × the price on the page. Above $100 the buying rate is scaled down with price, since cold visitors buy a $2,000 offer far less often than a $47 one.

Funnel typeVisitors who buyBasis
Sales page0.5%–2%0.5–2% of visitors buy
VSL0.5%–2%0.5–2% of visitors buy
Book a call0.2%–0.9%1–3% book × 20–30% close
Webinar registration0.6%–2.8%20–35% register × 3–8% buy
Webinar replay0.6%–2.8%20–35% register × 3–8% buy
Product launch0.6%–2.8%20–35% opt in × 3–8% buy
Opt-in0.3%–1.8%25–45% opt in × 1–4% buy the back-end offer
Squeeze page0.3%–1.8%25–45% opt in × 1–4% buy the back-end offer
Quiz0.3%–2%30–50% complete × 1–4% buy
Free + shipping1%–4%1–4% of visitors order
Tripwire1%–4%1–4% of visitors buy
Free trial0.3%–1.5%2–6% start a trial × 15–25% convert to paid
Ecommerce product page1%–3%1–3% of visitors buy
Other0.3%–1.5%0.3–1.5% of visitors buy

When no price is visible, the value per buyer comes from the offer’s ticket tier:

Ticket tierValue per buyer
Free offer$47–$497
Low-ticket$17–$97
Mid-ticket$197–$997
High-ticket$2k–$8k
Subscription$180–$900

An ad kept running for 60+ days is very likely paying for itself, so a funnel with such an ad still running is never estimated below its measured ad spend. That rule is a heuristic, and it’s labelled as one wherever it applies.

Winning score (0–100)

PartPointsFrom
Longevity30Days its longest ad has run (log scale, 365 days = full)
Scale20Ads running now (30 = full)
Testing10Ad versions while running (50 = full)
Reach20People reached per month, when disclosed (10M = full)
Revenue20Where its revenue estimate ranks among all funnels (percentile)

Rising = first ad started in the last 60 days and two or more ads already running. Trending = new ad creatives launched in the last 14 days.

How accurate is it?

Not measured yet. No verified numbers yet: accuracy is reported once owners verify funnels. Owners can verify their funnel from its page in the app with a read-only Stripe key; every verified funnel is compared with what we estimated for it, and the result is published here — including when we’re wrong.

What we never do

  • Show a number without a source: a missing input means a missing figure.
  • Present a modelled figure as an observed one — every estimate says which it is.
  • Submit forms, buy, or create accounts on the funnels we record: we only follow links.
  • Publish a funnel owner’s verified numbers without their opt-in.
Talk to a human