How to find out how a website makes money
Most websites earn from Google AdSense, a managed ad network such as Mediavine or Raptive, or affiliate links to stores such as Amazon. Each leaves public traces. This guide shows how to read them, by hand and with pubsleuth, and how far to trust what they say about a site's traffic and income.
Updated 64 questions in 11 chaptersAbout 41 minutes to read
By the pubsleuth team
Chapter 1
How websites make money, and the traces it leaves
Most content sites earn from Google AdSense, a managed ad network or affiliate links. Each of the three has to tell the company that pays it which site is which, and much of that is public. That is what pubsleuth reads.
What are the main ways a website makes money?
Link to this questionMost content sites earn in one or more of three ways. They show ads through Google AdSense, they hand their ad space to a managed ad network such as Mediavine, Raptive or Ezoic, or they earn a commission through affiliate links to stores such as Amazon. Many sites combine two of them.
- AdSense: Google's self-serve ad program. Google chooses the ads, and the site is paid for the clicks and views they get.
- Managed ad networks: companies such as Mediavine, Raptive and Ezoic that run a site's ads for it and share the revenue with it.
- Affiliate programs: a store, or a network of stores such as Awin, CJ or Impact, pays a commission when a reader follows a link and buys.
Other income, such as sponsored posts, memberships or a site's own products, leaves few public traces, so pubsleuth doesn't cover it.
What public traces does each kind of income leave?
Link to this questionEach one has to identify the site to the company that pays it, and much of that is public. AdSense needs a publisher ID in the page's ad code. Ad networks are named in the site's ads.txt file and in their own sellers.json. Affiliate links carry the member's tag or ID.
| Income | Where it shows | What pubsleuth reads |
|---|---|---|
| AdSense | The publisher ID in the page's ad code, in ads.txt and in Google's sellers.json | The page code, a public browser crawl and Google's sellers.json |
| Managed ad network | The network's lines in the site's ads.txt, its tag on the page, and its own sellers.json | Each network's sellers.json, checked against the site's ads.txt and page code |
| Affiliate program | The tag or member ID in links to the store, often behind a short link | The links on the site's home page and commercial pages, followed to where they lead |
These files and links exist so that advertisers, ad networks and stores can check whom they are paying. pubsleuth reads them as any visitor could and gathers them in one place, with the evidence for each finding.
How do AdSense, a managed ad network and affiliate links differ?
Link to this questionThey differ in who sells the ads and how the site is paid. With AdSense, Google sells the space and pays for clicks and views. A managed network sells the space on the site's behalf and shares the revenue. An affiliate program pays a commission only when a reader buys.
| AdSense | Managed ad network | Affiliate program | |
|---|---|---|---|
| Who sells the ads | Google, through its own auction | The network, to many ad buyers | No ads: the site recommends products |
| How the site is paid | For clicks and views | A share of the ad revenue | A commission on each sale |
| Public trace | The publisher ID in the page code | Lines in ads.txt and the network's tag | A tag or member ID in the links |
| pubsleuth service | AdSense Index | Network Scout | Affiliate Scout |
The difference matters when you compare sites. Two sites of the same size can earn very differently, depending on which of the three they use and how well it suits their audience.
Which pubsleuth service answers which question?
Link to this questionThere is one service for each kind of income. AdSense Index covers 819,808 AdSense sites with traffic and revenue estimates. Network Scout covers 173 ad networks and the sites verified on them. Affiliate Scout covers 89 affiliate programs and the sites that earn from them. The lookup box searches by any of their identifiers.
- Which other sites share this AdSense publisher ID? The reverse lookup.
- Which AdSense sites in a niche earn the most per visit? AdSense Index.
- Which sites run Mediavine, and which have left it? Network Scout.
- Which sites earn from Amazon Associates, and under which tag? Affiliate Scout.
Chapter 2
The reverse lookup: basics
What a reverse lookup is, what you can paste into it, what it shows for free and on a plan, and how far to trust what comes back.
What is a reverse AdSense lookup?
Link to this questionA reverse AdSense lookup takes an AdSense publisher ID, or a website, and lists the other sites that carry the same ID. Sites on one ID are paid through one AdSense account, so the list shows a publisher's other sites, with their niches and traffic.
- Paste an ID or a website into the lookup box.A pub- ID, a domain, an Amazon tag or an analytics ID all work.
- Read the list of sites.Each site comes with its niche and sub-niche, so you can see at a glance what the account publishes.
- Open a site.On a plan, its record shows traffic, a revenue estimate, its audience and the evidence that ties it to each ID.
pubsleuth runs the lookup against an index of 819,808 AdSense sites. Each one is filed by niche, audience country, visits and estimated revenue. So you can also start from a question instead of an ID: which sites in a niche earn most per visit, or which measured sites are growing.
What can you look up?
Link to this questionFour kinds of identifier: an AdSense publisher ID, a website address, an Amazon Associates tag or other affiliate ID, and a Google Analytics or Tag Manager ID. Each returns the sites in the index that carry the same identifier, so one known site leads you to the others.
| You paste | It looks like | You get |
|---|---|---|
| An AdSense publisher ID | pub- or ca-pub- and sixteen digits | The sites that its page code or Google's sellers.json ties to it |
| A website | example.com, or a full address | The site's own publisher ID first, then the other sites on it |
| An affiliate ID | An Amazon tag such as name-20, or a network's member ID | The sites whose affiliate links carry it |
| An analytics ID | G-, UA- or GTM- and its letters or digits | The other sites that load it, when two or more share it |
What does a lookup show for free, and on a plan?
Link to this questionSigned out, you can run 20 lookups a day and see up to 6 sites on an ID by name, with a count of the rest. A free account raises that to 50 lookups a day and up to 7 named sites. A plan names up to 500 sites a lookup.
| Signed out | Free account | Plan | |
|---|---|---|---|
| Lookups a day | 20 | 50 | No daily cap |
| Sites named on an ID | Up to 6 | Up to 7 | Up to 500 |
| Lists of sites to filter and sort | None | A preview, the first 5 of each view | Up to 10,000 sites a month in each service |
On a plan, each named site shows its monthly visits, estimated revenue, top audience country and how many sites its publisher has. A plan counts up to 10,000 different sites a month in each service, and a site you open again that month doesn't count twice. The pricing page lists each plan exactly.
How accurate is a reverse AdSense lookup?
Link to this questionThe links between sites and publisher IDs are findings, not guesses. A site is tied to an ID only when its own page code carries the ID, or when Google's sellers.json names the account for that site. An ads.txt line alone never counts, because ad networks list their own accounts there.
A site's record shows the evidence that ties it to each publisher ID, so you can check a surprising result yourself. Traffic and revenue are different: they are estimates, explained in How does pubsleuth estimate a site's AdSense revenue?
Chapter 3
AdSense publisher IDs
What an AdSense publisher ID looks like, how to find one on any website in a minute, and why one ID can turn up on sites that have nothing to do with each other.
What is an AdSense publisher ID?
Link to this questionAn AdSense publisher ID is the number Google gives each AdSense account: pub- followed by sixteen digits. In a page's ad code it appears with ca- in front, as ca-pub-. People also call it an AdSense ID or a pub ID. All the sites in one account carry the same ID.
The ID tells Google which account to pay for each ad shown, so it is written into the page for anyone who views its source. Google's help page Find your publisher ID shows the format.
How do you find a website's AdSense publisher ID?
Link to this questionOpen the site's ads.txt file by adding /ads.txt to its domain. Find the line that starts with google.com and ends in DIRECT: the number beginning pub- is usually the publisher ID. If the site has no ads.txt, view the page source and search for ca-pub-.
- Open example.com/ads.txt.Use the site's own domain in place of example.com. The file is plain text, one line per company.
- Find the google.com line that ends in DIRECT.The pub- number on it is usually the site's own publisher ID.
- Check the page source.Press Ctrl+U (Cmd+Option+U on a Mac) and search for ca-pub-. A match confirms the ID the ads use.
- Or paste the domain into the lookup.If the site is in the index, it shows the ID and the other sites on it in one step.
Lines that end in RESELLER usually belong to the ad networks that sell the site's space, not to the site. Checking the page source as well tells you which account the site's own ad code uses.
Chapter 4
ads.txt and sellers.json: the evidence
The two public files the ad industry uses to say who may sell a site's ads and who gets paid. They matter for AdSense and for every ad network, so it pays to know what each one proves.
What does an ads.txt line prove?
Link to this questionOnly that the site allows that company to sell its ad space. A DIRECT line names an account the site holds itself, and a RESELLER line names a company selling on its behalf. A line does not prove who holds a publisher ID, and lines often stay in the file after a site leaves a network.
The IAB Tech Lab's ads.txt standard defines the file. pubsleuth reads it to see which ad networks a site works with, and never to tie a site to an AdSense account on its own.
What is sellers.json, and how does pubsleuth use it?
Link to this questionsellers.json is a public file in which an ad system lists the sellers it pays, each with an ID, a type and, unless marked confidential, a name and a domain. Google publishes one for AdSense. pubsleuth checks each site's publisher ID against it, and marks an ID that Google lists as serving.
An ID that Google does not list is marked not serving. For ad networks, pubsleuth reads the sellers.json of each network we track that publishes one, to find the sites each network says it pays. The IAB Tech Lab's sellers.json standard defines the format.
How do ads.txt and sellers.json differ?
Link to this questionads.txt is published by a website and names who may sell its ad space. sellers.json is published by an ad system and names whom it pays. Read together, they let anyone check a chain: the site says a network may sell under an account, and the network says whose account it is.
- The site's ads.txtgoogle.com, pub-…, DIRECT: Google may sell this site's space under this account.
- Google's sellers.jsonSeller pub-… is a PUBLISHER, with the site's domain beside it.
- The chain checks outThe account belongs to the site, and Google pays it.
| ads.txt | sellers.json | |
|---|---|---|
| Published by | The website, at example.com/ads.txt | The ad system, such as Google or an ad network |
| Lists | The companies allowed to sell the site's ad space | The sellers the ad system pays |
| Relationship | DIRECT or RESELLER | PUBLISHER, INTERMEDIARY or BOTH |
| What pubsleuth reads it for | Which ad networks a site works with | Whether a publisher ID is serving, and which sites a network pays |
| Ties a site to an AdSense account? | No, never on its own | Yes, when Google names the account for that site |
How is a site confirmed on a publisher ID?
Link to this questionThere are three ways. Its page code carries the ID in the AdSense script, a public browser crawl saw the page ask for ads with that ID, or Google's sellers.json names the account for its domain. A site's record shows which one applied for each of its publisher IDs.
- Page code: the AdSense script with the ID is in the site's own HTML.
- Browser crawl: a public crawl, run in a browser, saw the page request ads with the ID.
- Google's sellers.json: Google lists the account with the site's domain.
Our crawler reads the page code itself. When a script adds the ad code later, the ID is missing from the raw HTML. So pubsleuth also reads a public crawl that loads popular pages in a browser and records their requests.
Chapter 5
What the index records about each of 819,808 AdSense sites
What a site is about, what kind of site it is, who visits and how many, whether its ads are serving and which sites share its publisher ID. Most sites carry most of these facts, and the data page shows how many have each.
How are a site's niche, sub-niche and cluster chosen?
Link to this questionA language model reads the site's homepage title, description and headings, and files it under a niche with a sub-niche in plain words, such as "unit converter". Sub-niches of the same kind are then grouped into clusters inside each niche, so sites that do the same job sit together.
- NicheThe broad market, such as Travel or Finance & Insurance.
- ClusterA group of sites that do the same job inside it, such as city guides.
- Sub-nicheWhat one site does, in its own plain words.
A cluster is the quickest competitor map for a topic: it holds the sites that do the same job, not only those on the same subject. The data page shows how many sites carry each fact.
What is a site type, and what does transactional mean?
Link to this questionA site type says what kind of site it is, apart from its topic, such as a blog, a news site, a tool, a store, a forum, a download portal or a game portal. Transactional sites are the ones people use to get a job done, such as tools, downloads and games.
A finance calculator and a finance blog are different businesses, with different traffic and different earnings per visit. The Transactional only switch keeps tools, calculators, converters, downloads, software and games, and drops blogs and news.
Where do a site's audience countries come from?
Link to this questionFor sites with measured traffic, from a third-party traffic panel. For the rest, from public per-country popularity data, which our model turns into a share of visits for each country. Each site also has a language, the one its pages are written in.
The country mix matters because advertisers pay very different amounts to reach different countries. A visit from the United States is usually worth several times one from India.
How are monthly visits measured or estimated?
Link to this questionFor sites in a third-party traffic panel, visits are measured. For the rest, our model estimates them from public popularity signals and the links between sites. Figures from the model carry a tilde (~), so you can tell an estimate from a measurement.
| Measured site | Modelled site | |
|---|---|---|
| Monthly visits | From the traffic panel | From our model, shown with ~ |
| Three-month trend | Yes | No |
| Audience countries | From the traffic panel | From public popularity data |
| Search keywords and volumes | Yes | No |
The model is fitted on the measured sites and cross-validated before it is stored. Only measured sites carry a three-month trend.
What does “serving ads” mean?
Link to this questionIt means Google's sellers.json lists the publisher ID found on the site, so the account is one that Google pays. A site whose ID Google does not list is marked not serving, even if AdSense code is still on its pages, and its estimated revenue is zero.
Set Ads on the page to serving to keep only the sites whose AdSense account Google pays, or to not serving to find the ones whose setup is broken or abandoned.
What is an exact-match domain?
Link to this questionA domain whose name is the keyword the site targets, such as an age calculator on a domain named for the calculation. pubsleuth marks a domain as exact-match when its name, letters only, is two or more words in a row from the site's own title, description, headings or sub-niche.
A brand written as one word doesn't count, and add-ons such as "online" or "free" are allowed around the keyword. Tick Exact-match domains (EMD) only to keep these sites.
Which sites show search keywords and volumes?
Link to this questionOnly sites with measured traffic. For them, the index shows their top search keywords and how many people search for each one a month, which tells you what readers come for and how large that demand is. Sites whose traffic comes from our model have none of these.
pubsleuth is not a keyword tool or a rank tracker. What it adds is which keywords already support earning sites.
How many of a publisher's other sites can you see?
Link to this questionA site's record shows how many other sites share its publisher ID and lists up to 50 of them, with their niches and traffic. The count tells you whether you are looking at one site or a portfolio, and how the publisher spreads its sites across niches.
The record opens with a plan, and a lookup of the publisher ID names up to 500 sites.
Chapter 6
How AdSense earnings are estimated
The method behind each revenue figure: visits, pages per visit and a page RPM for the niche and audience, shown as a range because every input is an estimate.
How does pubsleuth estimate a site's AdSense revenue?
Link to this questionIt multiplies the site's monthly visits by its pages per visit, then by a page RPM for its niche and audience countries, and divides by a thousand. The result is the middle of a range: the low end is 0.4 times the middle and the high end 2.5 times.
- Page viewsMonthly visits × pages per visit.
- Page RPMThe niche's base RPM for US traffic × the audience's country value × the keyword adjustment.
- Revenue rangePage views × page RPM ÷ 1,000, from 0.4 to 2.5 times that middle.
Pages per visit are measured where the traffic panel has them. Otherwise the model uses the median of the measured sites in the same niche. For measured sites, the page RPM is also scaled by what advertisers pay for the site's keywords against its niche's typical rate, to no less than half and no more than double.
What is page RPM, and why does the audience country change it?
Link to this questionPage RPM is what a site earns per thousand page views, as Google defines it. It depends on the niche, because advertisers pay more for some topics, and on the audience country, because they pay more to reach some readers.
| A visit from | Counts as |
|---|---|
| The United States | A full visit: the base the niche RPMs are set for |
| The UK, Canada, Australia and similar | About 80% of a US visit |
| Western Europe, Japan and similar | About half |
| Brazil, Mexico, Poland and similar | About a fifth |
| India, the Philippines and similar | About a tenth |
The model's base page RPMs for US traffic run from about $3 for games to about $21 for home, DIY and real estate. They are drawn from Google's own AdSense revenue calculator and from recent publisher benchmarks. They are the model's assumptions, not measurements.
Why is revenue shown as a range?
Link to this questionBecause every input is itself an estimate, and the model can't see what decides a site's real earnings: its ad density, ad formats, viewability, seasons and the account's own history. A site's true figure is typically between 0.4 and 2.5 times the middle estimate, so the range is wide on purpose.
Estimates are good for comparing sites and niches, and poor for stating what one site earned. The revenue filter and the Est. revenue sort use the middle of the range.
What does a site that isn't serving ads show?
Link to this questionAn estimated AdSense revenue of zero, because Google's sellers.json does not list its publisher ID. Next to the zero, the index shows what the same traffic could earn with a serving account, worked out the same way as for any other site.
That pair is how you spot sites with an audience and a broken or abandoned AdSense setup.
Chapter 7
Researching AdSense sites, step by step
How to narrow 819,808 AdSense sites to the ones you want: the filter and the sort for each question, five ready-made views and a worked example you can follow in ten minutes.
How do you find the AdSense sites in a niche?
Link to this questionOpen AdSense Index and choose the niche in the Niche filter, then narrow it to one kind of site with the Cluster filter. The count beside each choice shows how many sites hold it. Sort by Visits / month to see the largest first, or by Page RPM to see where a visit earns most.
- Open AdSense Index and choose a niche.The count beside each niche shows how crowded it is before you choose.
- Narrow it with the Cluster filter.A cluster keeps one kind of site, such as city guides within Travel.
- Sort the list.Visits / month puts the largest first; Page RPM puts first the sites where a visit is worth most.
How do you find sites with a US audience?
Link to this questionChoose United States in the Audience country filter. You get the sites that draw at least 30% of their visits from it. Raise Min. share from that country to keep only the sites that depend on it, or tick It's the #1 country to keep those where the US is the largest audience.
- Choose United States in Audience country.Sites with at least 30% of their visits from the US stay.
- Raise Min. share from that country, if you want.At 50%, only sites with half their audience in the US remain.
- Or tick It's the #1 country.That keeps the sites whose largest audience is the US, whatever its share.
How do you find growing sites?
Link to this questionSort by 3-month trend. The sites at the top are those whose measured visits rose most over recent months. Set a floor of Monthly visits, such as twenty thousand, so that small sites with a few extra visits don't crowd the top.
- Set a minimum in Monthly visits.Twenty thousand is a sensible floor; the Growing fast view uses it.
- Sort by 3-month trend.The fastest-rising measured sites come first.
- Add a niche or a country.To see where demand is moving in a market you care about.
How do you find sites with traffic but no working ads?
Link to this questionSet Ads on the page to not serving, then set a minimum of Monthly visits. What remains are sites with an audience whose publisher ID Google's sellers.json does not list. Their estimated revenue is zero, and each record shows what the traffic could earn with a serving account.
- Set Ads on the page to not serving.Only sites whose publisher ID Google's sellers.json doesn't list stay.
- Set a minimum in Monthly visits.So the list holds sites with a real audience.
- Open a record.Beside the zero, it shows what the same traffic could earn with ads serving.
How do you find exact-match tool sites?
Link to this questionTick Transactional only and Exact-match domains (EMD) only, then sort by Visits / month. That leaves the tools, calculators, converters, downloads and games whose domain is the keyword they target. Add a niche or a country to see how far the strategy goes in a market you are considering.
- Tick Transactional only.Tools, calculators, converters, downloads and games stay; blogs and news go.
- Tick Exact-match domains (EMD) only.Only domains named for their keyword remain.
- Sort by Visits / month.The most visited exact-match tools come first.
How do you find publishers with many sites?
Link to this questionSet Min. sites of the publisher to five, say, and sort by Publisher sites. You get the sites of accounts behind several AdSense sites, and the records show how each publisher spreads its sites across niches. Type a publisher ID into the Publisher ID filter to see one account's sites.
- Set Min. sites of the publisher.Five keeps accounts behind five or more AdSense sites.
- Sort by Publisher sites.The largest portfolios come first.
- Type one ID into Publisher ID.To see a single account's sites, which is the reverse lookup inside the list.
Which ready-made views can you start from?
Link to this questionAdSense Index has five quick views. Each is a set of filters and a sort that opens in one click, and you can change any of it from there.
| View | What it keeps | Sorted by |
|---|---|---|
| Top earners | Serving sites with at least $5k of estimated revenue a month | Est. revenue |
| Exact-match tools | Transactional sites whose domain is the keyword | Visits / month |
| US audience, 10k+ visits | Half or more of the visits from the US | Visits from the country |
| Publishers with 5+ sites | Accounts behind five or more AdSense sites | Publisher sites |
| Growing fast | Sites with 20k+ visits | 3-month trend |
Is a calculator site worth building? A worked example
Link to this questionHere is how to test an idea, a set of calculators, in about ten minutes. Each step narrows the same list in AdSense Index, and none of them is a prediction: together they show how many sites compete, how big the leaders are and what a visit is worth.
- Count the competition.Tick Transactional only and search for calculator. The count is how many calculator sites already exist.
- Keep the established ones.Choose United States in Audience country with a share of 50%, and set Monthly visits to at least ten thousand.
- Keep the earners.Set Ads on the page to serving and sort by Est. revenue. Read the top ten: their sub-niches show which calculators earn most.
- Check the momentum.Sort the same list by 3-month trend. Among the measured sites, those at the top are growing.
- Look behind the leaders.Open a few and look at the other sites on their publisher IDs. One publisher with dozens of calculators is a portfolio to compete with.
- Choose the niche.Compare the page RPM of the leaders in finance, health and construction, and pick the one where a visit is worth most.
Chapter 8
Managed ad networks: Mediavine, Raptive, Ezoic and more
Many established sites hand their ads to a managed network instead of running AdSense alone. Here is how to tell which network a site uses, how Network Scout verifies it, and how to follow sites as they join and leave.
What is a managed ad network?
Link to this questionA managed ad network runs a site's ads for it. It places the ad code, sells the space to many ad buyers at once and pays the site a share of the revenue. Mediavine, Raptive and Ezoic are well-known examples, and many smaller networks specialise in games, video or native ads.
Network Scout covers 173 ad networks, in groups such as ad management, gaming, contextual, native, video, supply platforms that work directly with publishers, and pop and push networks.
How do you check which ad network a website uses?
Link to this questionOpen the site's ads.txt and look for a MANAGERDOMAIN line naming the network, or DIRECT lines from the network's ad system. Then view the page source and search for the network's script. Network Scout does both for every site it tracks and shows the result in one row.
- Open example.com/ads.txt.A managed network usually adds dozens of lines at once, one for each of its ad partners.
- Look for a MANAGERDOMAIN line.It names the company that manages the site's ads. DIRECT lines from that network's ad system point the same way.
- Set the RESELLER lines aside.They name companies that resell the space and say little about which company manages the site's ads.
- Check the page source.Search for the network's name or script host. The tag on the page confirms the network serves the site now.
- Or look the site up in Network Scout.It shows each network with its evidence, and when the site joined or left.
How does Network Scout verify a site on a network?
Link to this questionIt starts from each network's sellers.json, which lists the sites the network says it pays. It then reads each site's own ads.txt and homepage. A site counts as verified when its ads.txt carries the network's DIRECT line or manager record, or its page loads the network's tag.
- The network's sellers.jsonLists the sites the network says it pays.
- The site's ads.txtA DIRECT line from the network's ad system, or a MANAGERDOMAIN record.
- The site's homepageThe network's ad tag loads on the page.
Each network on a site's row carries three evidence marks, sellers.json, ads.txt and page script, filled when that source confirms the network now. Sites also enter from lists and from the AdSense index, when their ads.txt has a DIRECT line from a network we track.
What do Verified and the other statuses mean?
Link to this questionVerified means the site's own ads.txt or homepage confirms a network. A site listed only in a network's sellers.json is not verified. Its status says why: its ads.txt has no network lines, it has no ads.txt, it blocked us, its domain is dead or unreachable, or it hasn't been checked yet.
| Status | What it means |
|---|---|
| Verified | The site's ads.txt or homepage script confirms the network |
| Not in ads.txt | In the network's sellers.json, but the site's ads.txt has no network lines: it has probably left or changed tier |
| No ads.txt | The site has no ads.txt file |
| Blocked us | The site answered with a bot wall |
| Domain dead | The domain no longer resolves |
| Unreachable | Timeouts, server errors, or a web page instead of ads.txt |
| Not checked yet | Its ads.txt hasn't been read yet |
How do you see when a site joined or left a network?
Link to this questionNetwork Scout compares each new version of a network's sellers.json with the one before, and each re-read of a site's ads.txt with the last. A site that appears is recorded as joined, one that disappears as left, and one that leaves one network and joins another as moved.
- Open the Changes tab in Network Scout.Choose a window: the last day, 7 days, 30 days, 90 days or year.
- Read the three tables.Joined a network, Moved between networks and Left a network, each with the date the change was detected.
- Narrow it to one network.The Network filter shows who joined or left Mediavine alone, for example.
- Or filter the site list by Added in sellers.json or ads.txt.The recent joiners, with their niche, traffic and revenue estimate.
What does Network Scout show about each site?
Link to this questionFor each site: its networks now and earlier, the evidence for each, its niche and type, its tool and download pages, and its estimated visits and ad revenue. It also lists the other sites that share its analytics, Tag Manager or AdSense IDs. The list starts with transactional sites, such as tools and downloads.
- Networks now and Earlier: the networks with live evidence, and those the site has left.
- Evidence: each network with its source, when it was first and last seen, and whether it is still active.
- Joins and leaves: every change detected, with the file it was seen in.
- Owner cluster: the sites that share its identifiers, and the IDs that tie them.
- Tool and download pages: the pages found through its homepage links and sitemaps, by keyword.
- Monthly visits and estimated ad revenue, both modelled and marked with ~.
The list's Show filter switches between transactional sites, the default, sites with three or more tool pages, and all network sites. A free account previews the first five rows of the site list; the Keywords and Changes tabs open with a plan.
How does Network Scout group sites that share identifiers?
Link to this questionIt groups sites that load the same Google Analytics, Tag Manager or AdSense ID, or that name the same owner domain in their ads.txt. A shared ID usually means one owner, but an agency or a developer can reuse one ID across clients, so each group lists the IDs that tie it for you to judge.
IDs found on very many sites are dropped as templates, as are IDs that span several networks or belong to a network itself. Paste a Google Analytics or Tag Manager ID into the lookup box to see the sites that share it.
How are a network site's traffic and revenue estimated?
Link to this questionVisits are measured where a third-party traffic panel has the site; otherwise they come from the AdSense index's traffic model and carry a tilde (~). Revenue multiplies the visits by a typical number of pages per visit and a rough page RPM for the site's niche or type, shown as a range from 0.4 to 2.5 times its middle.
What can you research with Network Scout?
Link to this questionWhich sites run a given network in a niche, which joined or left it recently, which networks sites like yours use, and which tools and downloads network sites build. Each question is one filter or one tab in Network Scout.
- Sites on one network in a niche.Choose the network in Network, then the niche in Niche / type, and sort by Est. visits / month.
- Recent joiners.Set Added in sellers.json or ads.txt to the last 30 days, then read their niches and sizes.
- Sites that left.Open Changes, choose a network and read Left a network, and Moved between networks for where they went.
- Tool and download ideas.Open Keywords: the tool and download pages found on network sites, grouped by keyword. A keyword many sites target is proven demand.
Chapter 9
Affiliate programs: Amazon Associates, Awin, CJ, Impact and more
Many content sites earn most of their money from affiliate links, not ads. Here is how to read a site's affiliate links by hand, how Affiliate Scout finds and judges them, and what its verdicts, portfolios and views show.
How do affiliate links work?
Link to this questionA site joins a store's program, such as Amazon Associates, or a network of programs, such as Awin, CJ or Impact. Each member gets a tag or ID that goes into every link it posts. When a reader follows the link and buys, the program credits that member with a commission.
Affiliate Scout tracks 89 affiliate programs: stores' own programs, networks of many merchants, and the link services and plugins that sit in between.
How do you find which affiliate programs a site uses?
Link to this questionOpen a review or comparison page, hover over the buy buttons and product links, and read where they lead. A tag= in an Amazon link or a network's ID in the address names the program. Short links such as /go/ or /recommends/ hide the destination until you follow them.
- Open a commercial page.A best-of list, a review or a comparison. Home pages often carry few affiliate links.
- Hover over the product links.The address in the corner of the browser shows where each one leads.
- Look for the member's ID.tag= in an Amazon link, or a network's publisher ID in an Awin, CJ or Impact link.
- Follow short and cloaked links.Links such as /go/ or /recommends/ redirect to the store; the final address carries the ID.
- Or look the site up in Affiliate Scout.It does all of this and lists the programs with the evidence for each.
How does Affiliate Scout read a site?
Link to this questionIt reads the home page and up to five commercial pages chosen from the sitemap, such as best-of lists, reviews and comparisons. It follows short and cloaked links to where they lead, matches each destination to a program, and stops once it has proof, so it reads only as much as it needs.
- Home page and sitemapPicks the pages most likely to carry affiliate links.
- Commercial pagesUp to five: best-of lists, reviews, comparisons.
- Links followedShort and cloaked links, up to five hops each.
- Programs matchedEach destination matched to a program, with the member's ID.
What do confirmed, likely and candidate mean?
Link to this questionThey say how strong the evidence is. Confirmed means at least one proven affiliate link, such as an Amazon link with a real tag. Likely means two strong signals but no proven link. Candidate means commercial markers only, or a single strong signal, and isn't counted as an affiliate site.
| Verdict | It takes | Examples of the evidence |
|---|---|---|
| Confirmed | At least one proof | An Amazon link with a real tag, a network link with a publisher ID, a short link that resolves to one, or an affiliate script such as Amazon's |
| Likely | Two strong signals | An affiliate disclosure, sponsored links, an affiliate plugin, several /go/-style links, a short link left unresolved |
| Candidate | Commercial markers, or one strong signal | Price buttons, product review markup, an "our pick", pros and cons |
| Mirror | A copy of another site | Kept so that the copy isn't counted twice |
How does Affiliate Scout group sites by affiliate ID?
Link to this questionSites that carry the same Amazon tag, network publisher ID or other member ID are grouped as one portfolio, because a member uses its ID across its sites. Media companies' portfolios are labelled corporate, and a tag copied from another site is marked copied and joins nothing.
IDs found on more than sixty sites are dropped as vendors' or templates' IDs. Paste an Amazon tag, the part that ends in -20, -21 and so on, into the lookup box to see the other sites that carry it.
What does Affiliate Scout show about sites and programs?
Link to this questionFor each site: its verdict and score, the programs found with their evidence, and the owner IDs it carries with its portfolio. It also shows the merchants and products it links to, its niche, and an estimate of traffic and revenue. Views across all sites show programs, merchants, products, owners and niches.
| View | What it shows |
|---|---|
| Sites | Every affiliate site with its verdict, programs, owner IDs, traffic and revenue band; confirmed sites by default |
| Programs | One row per program: the sites with proof, the sites with any evidence, and its share of confirmed sites |
| Merchants | The merchants the confirmed sites link to, and how concentrated their links are |
| Products | The most-promoted products: Amazon items and merchants' product pages |
| Owners | Portfolios tied together by shared affiliate IDs, corporate or independent |
| Niches | Each niche's traffic and revenue bands, with its top program and top merchant |
A free account previews the first five rows of the site list; the other views open with a plan.
How is an affiliate site's revenue estimated?
Link to this questionAn affiliate site's revenue is a rough band built from its visits. The model assumes a share of visits read commercial pages and a share of those click out. It multiplies that by the typical earnings per click of the site's programs and a factor for the audience country, from 0.3 to 3 times the middle.
It is uncalibrated: nothing in it has been checked against real affiliate income, so it is meant only to rank sites against each other. Sites without a traffic figure or without a proven link get no band.
What can you research with Affiliate Scout?
Link to this questionWhich sites in a niche earn from a given program, which programs and merchants sites like yours rely on, which products a niche promotes most, and which portfolios sit behind a set of sites. Each is a filter or a view in Affiliate Scout.
- Sites on one program.Choose it in Program, add a niche, and sort by Est. visits / month.
- Independent publishers only.Tick it to leave out media companies' portfolios.
- What a niche promotes.Open Merchants or Products and choose the niche.
- Who is behind a site.Open its record: the owner IDs it carries and the other sites in its portfolio.
Chapter 10
Who uses it, and for what
Five common jobs: mapping competitors, checking a niche before you build, checking a site before you buy it, choosing how to monetize, and studying a whole market. Each uses more than one service.
How do publishers use it for competitor research?
Link to this questionLook up your own site and note its cluster, then choose that cluster in the Cluster filter and sort by Visits / month. You get a ranked list of the sites that do the same job: their size, audience countries, estimated earnings and, for measured sites, their trend.
- Look up your own site.Note its niche and cluster on its record.
- Choose that cluster and sort by Visits / month.The sites that do the same job, ranked by size.
- Open the leaders.Their audience countries, page RPM and the other sites on their publisher IDs.
- Check how they earn beyond AdSense.Look them up in Network Scout and Affiliate Scout to see their ad networks and affiliate programs.
All of it comes from what sites publish in their own code and files: it is reading what rivals have made public, gathered in one place.
How do you check a niche before you build a site?
Link to this questionLook at three things: how many sites the niche has, what a visit to it is worth and how big its leaders are. The niche count shows how crowded it is. Page RPM and the country filter show the value of a visit, and the visits filters show the size of the leading sites.
| Signal | Argues for the niche | Argues against it |
|---|---|---|
| Number of sites | A few, with room between them | Many, all chasing the same readers |
| Page RPM of the leaders | High | Low |
| Audience countries | A strong share from the US, UK, Canada or Australia | Mostly low-paying countries |
| Size of the leaders | Modest sites that earn well | Only very large sites earn well |
| Trend of measured sites | Rising | Falling |
How does it help before you buy a website?
Link to this questionIt gives you an independent view to set beside the seller's figures. You see the site's traffic (and trend, if measured), audience countries, niche, page RPM, whether its ads are serving, its ad networks and affiliate programs, and the other sites on its publisher ID. A large gap is a reason to ask more questions.
- Compare the traffic.Set the seller's analytics beside the index's visits and, for measured sites, the trend.
- Check the ad setup.Is the AdSense account serving? Which ad networks does Network Scout find, verified or not?
- Check the affiliate income.Which programs does Affiliate Scout find on its pages, and how strong is the evidence?
- Look at the publisher's other sites.A portfolio seller may be selling the weakest site of several.
- Read the site.Check whether it reads like a made-for-advertising site, built for ads rather than readers.
How do you decide between AdSense, an ad network and affiliate links?
Link to this questionLook at what similar sites do. Find sites in your cluster, then check which of them run AdSense, which have moved to a managed network and which earn from affiliate programs. Sites of your size and audience show which model has worked for sites like yours.
- Find your neighbours in AdSense Index.Your niche and cluster, at a size close to yours.
- Check their ad networks in Network Scout.Which networks sites in that niche run, and which they left.
- Check their affiliate programs in Affiliate Scout.Which programs and merchants sites in that niche link to.
How do agencies and researchers use it?
Link to this questionTo study a whole market rather than one site at a time. A plan includes a CSV download of results for your own analysis, within a monthly allowance. The API looks up one domain, publisher ID or tag at a time, signed in with your account; an API plan is coming soon.
The acceptable use policy forbids reselling the data in bulk or building a competing index from it.
Chapter 11
Data, limits and privacy
Where the data comes from, which figures are findings and which are estimates, how fresh it is, what pubsleuth doesn't hold, how its crawler behaves, and the common ways to misread the results.
Where does pubsleuth's data come from?
Link to this questionMostly from public files and pages: the ads.txt files sites publish, the sellers.json files ad systems publish, the code and links on the pages themselves, and public web crawl data. Traffic comes from public popularity data and, for some sites, third-party traffic estimates.
| Source | What it gives |
|---|---|
| Sites' own pages | AdSense code, ad network tags and affiliate links |
| ads.txt files | Which ad networks a site lets sell its space |
| sellers.json files | Whom Google and each ad network pay |
| A public browser crawl | The ad requests pages make in a browser, when a script adds the ad code |
| Public popularity data | Popularity signals for the traffic model, and audience countries |
| Third-party traffic estimates | Measured visits, trend, countries and keywords for some sites |
What is measured and what is estimated?
Link to this questionWhich sites carry which publisher IDs, which ad networks they run and which affiliate programs they link to are findings, with the evidence in page code or public files. Traffic and revenue are estimates: revenue is shown as a range, and visits from our model carry a tilde.
| Fact | Kind | Where it comes from |
|---|---|---|
| Publisher ID on a site | Finding | Page code, a browser crawl or Google's sellers.json |
| Serving or not | Finding | Google's sellers.json |
| Ad network | Finding | The network's sellers.json, verified by the site's ads.txt or page code |
| Affiliate program | Finding | The links on the site's pages |
| Niche, sub-niche, site type | Classification | A language model reading the site's own pages |
| Monthly visits, audience countries | Measured or modelled | A third-party traffic panel, or our model (~) |
| Trend, keywords | Measured | The traffic panel, for measured sites only |
| Revenue, page RPM | Estimate | Our model, from visits, niche and audience countries |
How fresh is the data?
Link to this questionEach service's findings reach the website once a day, and the niche table on the home page shows the date its counts were taken. A single site may have been checked earlier than that, and sites change: ads are removed, domains are sold and networks are switched.
Network Scout re-reads sellers.json files daily and each site's ads.txt on a rolling cycle. The affiliate crawler works through its queue in runs, so an affiliate site's First seen and Last checked dates say how recent its verdict is.
How does pubsleuth's crawler behave, and can a site opt out?
Link to this questionIt reads the public files and pages that sites publish for the ad industry and for every visitor, and identifies itself as PubSleuthBot. A site that blocks PubSleuthBot in its robots.txt stops the ad network and affiliate crawlers reading its pages. They wait between requests to the same site and don't get around bot walls.
Files published for the ad industry, such as ads.txt and sellers.json, are still read, since that is what they are for. The AdSense checks don't read robots.txt. They read a site's ads.txt, its home page and up to two linked pages. They also open the home page and at most one article in a browser that stops every ad request before it reaches Google. For anything else about a site you run, write to us through the contact page.
What does pubsleuth not hold?
Link to this questionIt holds no email addresses, phone numbers or other contact details, and it has no outreach, emailing or scraping tools. It is a research tool about how websites make money, not a lead list. It does not show Domain Authority or other third-party SEO scores, and it is not a keyword tool or a rank tracker.
It is not affiliated with Google, any ad network or any affiliate program. The acceptable use policy forbids spam, harassing site owners and reselling the data in bulk.
What are the common ways to misread the results?
Link to this questionFive mistakes come up most often. Here they are, each with the habit that avoids it.
| Mistake | The habit that avoids it |
|---|---|
| Treating an estimate as a measurement | Quote revenue as a range, and never present an estimate to someone else as a fact. |
| Treating a shared ID as proof of one owner | Take it as a strong clue, and check the sites before you conclude anything. |
| Comparing revenue without the audience country | Two sites with the same visits can earn several times apart, so compare like with like. |
| Reading an ads.txt line as proof | A line says who may sell the space, not who does. Look for verification by page code. |
| Studying only the biggest sites | If you are starting small, filter to the size you can reach and learn from those sites. |
Look up a publisher ID
A publisher ID, a website, an Amazon tag or an analytics ID.