gtmcanon[1]

#Methodology

How gtmcanon evaluates software, what counts as evidence, and what gets a tool excluded. Versioned, dated, and changed in public.

Version v0.1 · Effective 2026-05-18 · Re-verification cadence 90 days

#The standard

Verifiability. Every recommendation in a gtmcanon review points at evidence the reader can open in a browser: a permalinked operator quote, a vendor's own documentation, or testing notes we logged on the day. If a claim can't be sourced that way, it doesn't run. We don't promise to cite it later; we just don't print it.

#The six criteria

We score each tool against six criteria. Per-criterion weights ship in a separate rubric file alongside the next version of this page, and the scores themselves sit next to the final verdict on every review. A reader can argue with the verdict piece by piece instead of either trusting or dismissing a single number.

  1. Operator-validated fit. Does the tool actually do what the vendor claims, at the team size and ICP they claim it for? Reddit, X, LinkedIn, and operator Slack communities with public web archives are where we look. Anonymous quotes don't count.
  2. Methodology transparency. Does the vendor publish how its data is sourced, how its models are trained, and how its rankings are produced? Black-box vendors lose points in proportion to the opacity. Saying "proprietary" is not an answer.
  3. Pricing legibility. Can a buyer find out what this costs without a sales call? Demos-to-see-pricing is a deduction. So is publishing a sticker price and burying the real cost in seat fees and platform fees and per-model add-ons.
  4. Integration health. Do the documented integrations work as advertised, at the documented scale, with the documented vendors? An integration listed on a homepage but broken in production scores worse than no integration claimed.
  5. Vendor responsiveness. Does the vendor reply to operator-reported issues in public, with a named engineer or product manager, on a measurable cadence? Silence in operator forums is a signal.
  6. Total cost of ownership. What does the tool cost beyond the line item: admin time, training, integration debt, switching cost when you leave? The published price is the floor; the rubric asks for the ceiling.

#How we source evidence

Every claim in a gtmcanon review comes from one of three places:

  1. Permalinked operator quotes from Reddit, X, LinkedIn, named blogs, or operator Slack communities with public web archives. We cite inline with handle, platform, and ISO date so the reader can verify in a click.
  2. First-hand testing by the gtmcanon team, with notes logged on the day of testing.
  3. Vendor documentation or public statements, quoted with URL.

If a claim doesn't fit one of those three, we leave it out. That's an editorial cost we eat on purpose.

#How we update

Every review records the methodology version it was scored against. When the rubric changes (to add a criterion, to reweight, to tighten a definition), the change ships first in the public rubric repository and only after that on the site. That lag is the audit trail. It's deliberate, and it's occasionally inconvenient.

Every published review is dated and last-verified-dated. When pricing, features, or vendor status change, we update the review and log the revision in the history block at the bottom of the page. Old facts don't get quietly overwritten.

#Disqualification rules

A tool is excluded from coverage only on the basis of criteria in this rubric. Affiliate status is not one of them. A tool without an affiliate program is evaluated on the same standard as a tool with one. Tools that pay gtmcanon $0 will be top-ranked when the methodology supports it. That is the entire promise the brand exists to make.

#Where this fails

Methodologies break. This one will. The honest list:

  • A vendor responds well to public scrutiny precisely because they know we're watching, and stops responding the moment we move on.
  • Operator quotes get gamed by vendors quietly seeding sympathetic practitioner accounts.
  • A category matures past where the rubric was designed, and the criteria stop separating winners from losers.

When that happens, the rubric versions. When it happens publicly, the rubric versions and the affected reviews log it. The methodology is only as good as the next version of itself.

#What this is not

This is not a Magic Quadrant or a Wave, and it isn't anyone's annual paid ranking. gtmcanon doesn't produce sponsored content, vendor case studies, or paid leaderboards. The category we cover is structurally hostile to those formats; the brand only exists because it refuses them.

#Versions

| Version | Effective | Changes | |---|---|---| | v0.1 | 2026-05-18 | First published. Six criteria, no weights yet. Weights ship in v0.2. |

The full rubric document, with per-criterion weights and scoring procedures, ships as a public repository alongside v0.2. The methodology applied to the published Peec AI review is the worked example.