How We Compare Tools: Facts, Buyer Checks and Review Trends
We count public reviews and analyse a recent sample. We don't test tools.Every tool gets the same ten buyer checks, read on the vendor's own pages and dated, and the public reviews we analyse are sorted into fixed themes.
The same ten checks for every tool
Ten things a buyer wants to know before paying, each read on the vendor's own pages. A review shows all ten, except a fact we had to leave out (see below); a list row shows up to five ✓.
- Firm monthly priceA price per month in dollars, billed monthly, is published: not a quote, not annual only
- Free planA free plan is listed on the pricing page
- Free trialA free trial is published
- No card to tryThe vendor writes that no credit card is needed to start
- Pay month to monthA month-to-month option is shown, not annual only
- Refund possibleThe published terms allow a refund, in full or in part
- SOC 2 statedThe vendor states a SOC 2 audit: an outside audit report on how it protects your data
- Company login (SSO)You can log in with your company account (single sign-on), on some plan
- Connects to your CRMA native HubSpot, Salesforce, Pipedrive, Zoho or Dynamics integration is listed
- Live supportChat or phone support is published, not email only
- published by the vendor
- not offered, per the vendor's pages
- not published, or not read on its pages
We never total a tool's checks: a total would be a score, and these ten do not matter equally to every buyer. A – is not a ✕: the vendor may offer it without publishing it, or we may not have read it.
Where every fact comes from
Five steps, the same for every tool. Pages print the result and its date; the proof stays in our records.
How we read public reviews
Reviews give us trends: no quote, no score, no star rating, and no review site named.
- Public reviewsThe totals published by the review sites we count, added up, rounded down and dated. We leave out a review page shared with another product, and a page the review site itself has penalised.
- Reviews analysedRecent reviews, read by an AI model and sorted into fixed themes; a second model re-reads a sample of each tool's reviews. Left out: reviews by non-users, about another product, or by someone the review site marks as tied to the vendor, and, for a tool that does many things, reviews about other uses.
- MentionsHow often a topic comes up, counted apart for praise and criticism. For most tools on this site, we add the topic counts one review site publishes to the counts from the reviews we analysed on other sites. Topics are listed in order of how often they come up.
How lists are ordered, and what each one tells you
Four ways to read a list: what each one tells you, and what it does not.
Most public reviews first, then by name, on every comparison list and directory, with the count on each row. Alternatives pages put their cards in the same order.
- It tells you
- How many public reviews a tool has.
- It does not tell you
- That it suits you, or that it beats the next tool. A count favours big vendors.
Type a feature, tick what you must have, and sort by most public reviews, lowest firm price or name.
- It tells you
- Which tools show what you asked for, as each vendor lists it.
- It does not tell you
- That a tool lacks a feature: we show only what we found. Typing hides the tools whose name and listed features do not contain your words; a tool that misses a tick drops below the line, never off the list.
On each comparison list: lowest firm monthly price with a free trial, most public reviews, most favourable mentions of its most praised theme. No tool gets two cards: a card skips a tool already shown and says so.
- It tells you
- Which tool meets the rule printed on its card.
- It does not tell you
- That it is the right choice. Prices are compared as published, per seat or per workspace, and mentions favour big vendors.
One to three plain labels under a tool's name, from a closed list, each tied to a fact read on the vendor's pages.
- It tells you
- What a tool is for, at a glance.
- It does not tell you
- How well it does that, or that a tool without a label cannot do it.
Compare by category: the pages of each kind of tool
Each row links comparison pages of that category. A tool in several categories is counted under its main one.
What pages do, and what none does
Rules that scripts check before publication. A page that fails a check on its data is not built.
- Print the unit next to every review count: public reviews, reviews analysed, favourable reviews or mentions.
- Print the date a price was read on the vendor's page, and flag a reading more than six months old.
- Where a page shows buyer checks, take them from the same list of ten, with a date behind every ✓ and ✕.
- Close every page that has an affiliate link with the affiliate disclosure and its link.
- Sort comparison lists and directories by public reviews, then by name, with the count on each row.
- Quote a review, or reproduce a platform's star rating.
- Print a score, a rank or a total of the ten checks.
- Present a count as a share of users: pages count reviews and mentions, not customers.
- Name a review platform, anywhere on the site.
- Claim the site has used a tool itself.
- Put an award or a word of approval by a vendor's name, except one Editor's pick: Skylead, on two LinkedIn lists.
Who pays us, and what the order reads
Some vendors pay us a commission when you subscribe through a Try or Choose button. Which tools get their own alternatives page was decided partly on our affiliate programmes, alongside search demand.
- Comparison lists: public reviews, then nameOur data records which tools have an affiliate link. The order does not use that information: a script checks it before publication.
- The tools a list showsA second script switches that information for every tool (tools with a link lose it, the others get one), rebuilds every page and checks that the tools shown and their order stay the same.
- The disclosure closes the pageOn every page with an affiliate link, the affiliate disclosure and its link close the last card.
What this method can't tell you
Facts and counts have limits. Four things this method cannot do.
A price or a plan can change the day after we read it. Each page prints when its facts were read: check the vendor's page before you pay.
Trends count what reviewers raise. Reviews describe other people's accounts: your domains, your list and your copy are not in them.
A big vendor usually collects more reviews and more mentions than a small one. The default order shows that count, not which tool is better.
Reviews come only from users who chose to write one, and some were invited by the vendor or offered a reward. We keep invited reviews; each page prints the counts a trend rests on.
What changed, and why
Changes to the method, with their date and the reason.
- Scores out of 10 removed: pages now print facts, ten buyer checks, review counts and trends.A single number needs weights chosen by the site; facts, dates and counts do not.
About the method
The site does not test tools. It counts public reviews and reads prices on vendors' pages, so each price comes with the date it was read and each trend with its count.
The default order reads the total of public reviews, then the name, so a vendor with more public reviews sits higher. The order does not use which tools have an affiliate link: a script checks it before publication. A vendor can report a wrong price or limit through the contact page.
The vendor does not publish that fact in that form, or we did not read it on its pages. It is not a no: a ✕ means not offered, per the vendor's pages.
A trend needs enough reviews analysed and enough mentions. Below that, the page says too few for a trend; with no review analysed, it shows facts only.
Some vendors publish prices in euros, Canadian dollars or pounds sterling; the ≈ figures are our conversion at the European Central Bank euro reference rates of the date shown, and never enter a comparison.
Pages are assembled by script from dated records. Facts are read and texts written with AI, then checked by script before they are validated, and no page is built from a record that has not been validated.