Methodology v4.5
How we rank 12,100+ software tools
Rankings built from public review volume and rating (G2, Capterra, Trustpilot, app stores) and media mentions, plus a public worked example so you can see exactly which signals move a tool up. This is the version of the methodology you can hold us to.
The short version
- 1. Category rankings are a composite score built from two observable signals: public review volume and rating (G2, Capterra, Trustpilot, app stores), and media mentions.
- 2. That composite is a sort key, not a star rating. We never show it as a /100 on tool pages because a private internal number with no transparent review corpus behind it would be misleading.
- 3. A separate profile-completeness score gates indexing: tools with too little verified data are deindexed and kept out of guide top picks.
- 4. What you do see on tool pages: the external rating aggregated from G2, Capterra, App Store, and Play Store, plus our written editor's take and the moderated community reviews.
- 5. Sponsorships and vendor outreach never change a tool's ranking score. Some offers on our pages may be paid placements or contain affiliate links. A sponsored box is labelled and sits outside the ranking. A paid guide position can affect displayed order without changing that score. The full list of things that do not count is in section 6.
- 6. The "by the numbers" table on a tool page is a different thing: dated, first-party positions among tools in the same category, described in section 7.
1. Two numbers you can see, one you cannot
Toolradar surfaces two numbers on tool pages: an external rating aggregated from public review sites, and a community rating from moderated Toolradar reviews. Behind the scenes there is also an internal composite ranking we use to sort lists, plus a completeness score that gates which tools are eligible to appear in guides. Those internal numbers are intentionally private.
External rating
The number shown on every tool page. Aggregated from G2, Capterra, Trustpilot, the Apple App Store, the Google Play Store, and the Chrome Web Store where available. When several sources are available, the rating is their average weighted by the logarithm of each source's review count (a source with 10,000 reviews weighs about twice as much as one with 100). We show the source and review count next to it so you can verify.
- Backed by a public review corpus
- Source named on the tool page
Community rating
Calculated only from Toolradar reviews approved by a moderator, including those written after a reader claims the free buyer's guide. Each review is email-confirmed, manually moderated, and flagged if it shows the patterns we associate with review bombing or vendor astroturf.
- Moderated before publication
- Updates as new reviews land
The two pools never mix
A Toolradar review never enters the G2, Capterra, or Trustpilot aggregate, and an external rating never enters the community rating. The guide-for-review prompt (the lead magnet on tool pages) feeds the community pool only, after the same email confirmation and manual moderation as any other Toolradar review. When a page shows "4.8 (880)" next to a G2 or Capterra label, that number is the public-site snapshot, dated, and does not include our first-party reviews.
Internal ranking and completeness scores (not shown on tool pages)
The composite ranking decides which tools sort highest in a category and which qualify for the editor's top picks in a guide. A separate 0 to 100 completeness score decides which tools are indexed at all: profiles too thin to vet in good faith get deindexed. Showing either as a star rating on tool pages would imply a precision we cannot defend without a public review corpus behind every digit, so we keep them backstage and use the external rating in the UI instead.
2. How a category ranking is built
The composite ranking that orders a category combines two signals that are both observable in the wild, not editor opinion. Review volume and media counts are log-scaled so a handful of tools with huge review corpora do not flatten everything else, and so a tool with thirteen reviews cannot outrank an established brand on volume alone.
External review strength
Public review volume from G2, Capterra, Trustpilot, and the app stores, multiplied by the aggregate star rating. Volume-weighted on a log scale: a 4.5 from tens of thousands of reviewers contributes far more than a 4.9 from a dozen. We normalize every source to a real 0 to 5 scale before it counts, so a scraper hiccup cannot inflate a small tool.
Media momentum
Stories about the tool in independent press, blogs and newsletters. A vendor's own blog, review sites, forums and sites that republish other outlets do not count, and a story carried by several outlets counts once. Each outlet has a tier: a tier 1 story (major news outlet) weighs 3, tier 2 (established tech or trade outlet) 2, tier 3 1, and half as much when the tool is only named in the body. Within an outlet the best story counts in full, the next one half, the third a third, so volume from one site cannot swamp the list. The last 30 days count twice on top of the all-time total, and both are log-scaled. Currently 15,600+ stories from 330+ outlets.
Profile completeness (indexing gate, separate from ranking)
A separate 0 to 100 score for whether the Toolradar profile has verified pricing, a written editor's take, structured pros and cons, categories assigned, a working website, and a real logo. It does not push a tool up the ranking. It decides eligibility: profiles too thin to vet in good faith are deindexed and left out of guide top picks.
What the ranking decides
Deindexed tools are excluded from category sitemaps and guide consideration. They remain accessible via direct URL so the data we have stays public.
3. A worked example: monday.com
monday.com: Organize work, automate tasks, and connect tools for team clarity. Here is how each ranking signal reads today, from the same public inputs shown on its Toolradar profile. We keep the raw internal number backstage (see why above) and describe the evidence instead.
4. Review verification
The community rating shown on tool pages is computed only from reviews that a moderator approved, and a moderator can only approve a review whose author has confirmed an email address (a rule enforced since October 2026). The other checks below help the moderator decide. We would rather have ten honest reviews than a thousand we cannot stand behind.
Email confirmation
Every reviewer must verify a working inbox before the review goes live.
Identity verification
A reviewer who signs in with LinkedIn gets a Verified label, with their photo and, if they gave one, their LinkedIn link, next to their review. A confirmed email shows as Email confirmed. Reviews written from an email address on the product's own website domain are not published.
Manual moderation
Every approved review has been read by the editor or a designated moderator before publication.
Pattern detection
Bursts of reviews on one tool, near-identical text, and many reviews from one IP or network are flagged for the moderator.
Reader reports
Any reader can flag a review. Flagged reviews get re-moderated within 48 hours, the editor signs off on the outcome.
5. How often things get refreshed
Weekly for the top 500 tools by traffic, otherwise on every guide rebuild or when a vendor flags a change.
Ingested daily from Dupple's media monitoring. Ranking scores are recomputed every week from all-time and last-30-day stories; trending is recomputed every day from the last 7 days.
Rolling refresh, oldest snapshot first, on tools with enough editorial standing to be cited. Each public rating on a listicle shows the snapshot date. A frozen count is a frozen citation.
Refreshed when a vendor ships a major release, when reader pushback forces a rethink, or on the annual category sweep.
Reviewed quarterly. 400+ categories live today, pruned when overlap gets confusing for readers.
This page is reviewed every quarter. Material changes get a version bump at the top and a note in the changelog.
6. What does not influence a ranking
- × Sponsorship in the Dupple newsletter network. A tool sponsoring Techpresso or any other Dupple newsletter does not get a ranking bump. The order stays put.
- × Paid placements. They do not change a tool's ranking score, which is computed only from public review volume and rating (G2, Capterra, Trustpilot, app stores) and media mentions. Some offers on our pages may be paid placements or contain affiliate links.
- × Vendor outreach. Emails asking us to "take another look" or PRs pitching a new feature do not move a ranking. We use those as input for the next editor pass and that is it.
- × Affiliate revenue. Some outbound links are tracked for analytics. None of them pay us per click in a way that would shift a ranking. Our full revenue breakdown lives at /how-we-make-money.
- × Social media noise. A viral tweet does not produce a ranking change. A measurable spike in tracked press coverage does, through the media-momentum signal, on its own merits.
- × Editor sympathy. If the editor likes a vendor personally, that does not move the order. The signals are the signals. Personal opinion goes in the editor's take where readers can see it and argue with it.
7. Toolradar Data: the numbers on tool pages
The "by the numbers" table on a tool page, the leaderboards on a category's statistics page and the data returned by our API and MCP server all come from one weekly snapshot. Each row says what it measures, which tools it is compared with and when it was computed. It is separate from the category ranking composite above and never feeds it.
- Peer group. Every position is among published tools of one category. When a tool sits in several categories, the table uses the most specific one that has enough data.
- Minimum sizes. A position is shown only when at least 8 tools in the group have data for that row. A "#2 of 34" badge needs a group of at least 15 tools and a place on the podium (or the top tenth for groups of 20 or more). Shares of a group are written as counts such as "8 of 30" until the group has 50 tools; growth is a rate and is shown as a percentage next to its base.
- Position 1 is best. For price, the lowest entry price ranks first. Ties share the best rank and are marked "joint"; a badge is given only if the whole tied group is still on the podium. Under each position the table says who it is counted among: tools with 20+ reviews for a rating, tools with 50+ visits for visits, and so on.
- Prices. Only tools whose price we read on the vendor's own pricing page are counted, and a tool is compared only with tools on the same basis: the same currency and the same thing priced (a flat team plan, the same per-seat unit such as user or agent, or no unit stated).
- Traffic rows. Visits count distinct visitor sessions in our first-party analytics, with bots excluded. They count visits from every source, including visits that arrived through a sponsored slot.
- What is not an input. Advertising spend, campaigns, clicks sold to advertisers and placement fees are not read by the computation.
- Freshness. Recomputed weekly. The "as of" month on every table is the month of the last computation, and a row left over from a missed run is never shown next to fresh ones. A rank history restarts when a tool drops out of a metric's group.
| Row | What it measures | Source | A tool needs |
|---|---|---|---|
| Entry paid plan | Monthly-equivalent price of the cheapest standard paid plan (annual prices divided by 12). A tool is compared only with tools priced in the same currency and on the same basis: a flat team plan, the same per-seat unit (per user, per agent...), or no unit stated. | Prices read on each vendor's own pricing page; only tools whose price we verified are counted. | A verified, parseable paid plan. |
| Free plan | Whether the tool has a permanently free plan (free or freemium pricing model). | The pricing model recorded on each Toolradar profile. | None. |
| Integrations listed | Number of integrations recorded for the tool on Toolradar. | Toolradar integration records. | At least one integration recorded. |
| Features tracked | Number of standard features (API access, SSO, mobile app and so on) recorded for the tool. | Toolradar feature records. | At least one feature recorded. |
| Listed as an alternative to | Number of other published tools whose Toolradar alternatives list includes this tool. A measure of how often a tool is considered a real substitute. | The Toolradar alternatives graph. | Listed as an alternative at least once. |
| User rating | The rating shown on the tool's page: Toolradar reviews once there are enough, otherwise the aggregate of public review sites. Only tools with at least 20 reviews are ranked against each other. | Toolradar reviews and review-site aggregates, each with its own source and count on the tool page. | At least 20 reviews behind the rating. |
| Media mentions, last 30 days | Number of distinct articles published in the last 30 days by independent press, blogs and newsletters that mention the tool. A story carried by several outlets counts once; the vendor's own blog, review sites, forums and republishing sites do not count. | Dupple's media monitoring (press, blogs, newsletters), counted from the article dates at each computation. | At least one mention in the window. |
| Toolradar page visits, last 30 days | Distinct visitor sessions on the tool's Toolradar page over the last 30 days, with bots excluded by our analytics. A session counts once however many times it reloads the page. | Toolradar's first-party analytics. | At least 50 visits in the window. |
| Visit growth vs previous 30 days | Change in visitor sessions between the last 30 days and the 30 days before. | Toolradar's first-party analytics. | At least 100 visits in each window. |
| Compare-page visits, last 30 days | Distinct visitor sessions on head-to-head comparison pages that include the tool, over the last 30 days. | Toolradar's first-party analytics. | At least 20 compare-page visits in the window. |
The category data is published under CC BY 4.0 with a CSV on each category statistics page. See something wrong in a row? Tell us.
8. Methodology changelog
Media momentum now counts stories, not raw matches: only independent press, blogs and newsletters (no vendor blogs, review sites, forums, republishing sites or Dupple's own newsletters), no mention of a tool on its own website, nothing for unpublished tools, and one story carried by several outlets counted once. Each story is weighted by outlet tier (3 / 2 / 1), half when the tool is only named in the body, with less weight for each further story from the same outlet. Trending press points are log-scaled instead of a fixed number per article. The worked example now reads live data.
Rewrote the review verification section to describe how reviews are checked today: a confirmed email address and moderator approval for every review, a Verified label only for a LinkedIn sign-in or a check by Toolradar, no publication of reviews written from an email address on a product's own website domain, bursts, copied text and shared IPs flagged for moderators, and reports from readers and vendors. Removed a stated account-age rule that was never applied.
Corrected the sentences on this page and on the About page that named only G2 and Capterra: the review signal counts every public review source on file for a tool (G2, Capterra, Trustpilot and the app stores), as the formula section already said. The formula is unchanged.
Added Toolradar Data (section 7): dated positions among tools in the same category on tool pages, category leaderboards and the API. The category ranking composite is unchanged.
Removed the featured signal (Featured in Techpresso) from the ranking score. It no longer adds anything to a tool's position. The score now uses two signals only: G2/Capterra review volume and rating, and media mentions. The Featured in Techpresso badge and page are unchanged.
Documented that Toolradar reviews and G2/Capterra aggregates are two pools that never mix, including reviews written after the buyer's-guide prompt. External ratings now refresh on a rolling oldest-first cadence and listicles show the snapshot date next to the count.
Rewrote this page to describe the ranking exactly as the code computes it: a composite of external review strength, media momentum, and editor-featured picks, with profile completeness as a separate indexing gate. Removed the fixed-percentage editorial rubric, which no longer matched the live system.
Replaced the import-time editorial score as the primary category sort with the composite ranking, because the old score was set once at import and routinely promoted obscure tools above well-known brands. Introduced the media-momentum signal backed by the tracked source network.
Split community rating from external rating. Added profile completeness so tools with thin data stop appearing in guide top picks.
Toolradar launches with an import-time editorial score and aggregated external ratings.
Questions about the ranking
How does Toolradar rank tools within a category?
Why don't you display an internal score on tool pages?
How does Toolradar verify reviews?
What does NOT influence a tool's ranking?
Think a ranking is wrong? Tell us why.
Vendors who spot a stale price, a frozen review count, or a fact we got wrong: use the official form. Bring the URL plus a primary source. We correct the page; exclusive first-party facts you want models to read belong there too.