Methodology

Pricing and features

Every price and feature fact is read from the vendor's own pricing or documentation page and recorded with the date it was verified. Nothing is copied from comparison roundups, because those go stale quickly and errors propagate between them unchecked.

Where a vendor does not publish a figure, the field is left empty rather than estimated. You will see “not stated” or a blank instead of a plausible-looking guess.

Vendors quote headline prices on different bases — some show the monthly rate, others the discounted annual equivalent — so every price carries a note saying which. A table that stacks a monthly figure against an annual one without saying so makes one product look cheaper than it is. Where we have confirmed both, both are shown.

Keeping it current

A verification date is only worth something if it is real, so each tool carries its own rather than the site sharing one. Pages listing several tools show the range across them.

Two automated checks run against this catalogue. The first asks whether each price we publish still appears on the vendor's own page, and flags it for a human to re-read if it has vanished — it never rewrites a figure by itself, because a tool that guesses at prices would defeat the point of the exercise. The second checks our own writing against our own data, catching a review that still quotes a price we have since corrected. Several vendors block automated reading or render prices in JavaScript; those are checked by hand.

Ratings

Ratings out of five are editorial judgements, not averages of user reviews. They reflect how well a tool delivers on what it claims to be, at the price it charges, for the audience it targets.

This is why our structured data marks them as a review by a named author rather than as an aggregate rating. Presenting an editorial score as though thousands of users had voted would be a straightforward misrepresentation, and it is one of the more common forms of dishonesty in this category.

Field tests, and why most say “pending”

Deliverability and reply rates are the numbers buyers most want and the ones most often fabricated. A figure like “97% inbox placement” is meaningless without knowing the domain age, warmup period, list quality, send volume and seed accounts behind it.

So the rule here is simple: we do not publish a performance number we have not measured ourselves. Comparison pages show a field test section marked pending with the methodology we intend to use. It stays pending until the test is actually run. An empty result is honest; a borrowed one is not.

When tests are published, they will state the sending domain age, warmup schedule, list source and verification method, sample size, seed accounts and date — enough for you to judge whether the result transfers to your situation.

How recommendations are made

Our tool finder weights your answers against verified catalogue facts — price, free tier, mailbox limits, API access, database size, multichannel support. Commission rates are not an input. The lowest-paying tool in our catalogue wins the “I need a free option” path outright, because it is the only one with a genuine free tier.

The result screen lists which of your answers produced the match, so you can audit the reasoning rather than trust a score.

Corrections

Pricing changes constantly and we will sometimes be behind. If a figure here is wrong, email matt@toolcloud.ai with a link to the current page and it will be corrected. Vendors are welcome to submit corrections; doing so does not buy a better verdict.

When we get something wrong we say so on the page rather than quietly editing it. If you read a review last month and it said something different, you should be able to see that it changed and why — a correction you cannot see is not much of a correction.