Skip to content

Methodology

The advisor is only worth something if you can check how it decides. This is what we apply, and the rules version every plan is computed with.

Editorial principles

  1. Declared neutrality

    We include Trusteed's competitors when they are the best option for the case.

  2. Disclosure

    Every Trusteed product profile carries the «Our product» label. The recommendation engine gives it no extra weight.

  3. Honest status

    Every profile and protocol states its status (stable, beta, draft specification or deprecated) with the date it was last verified.

  4. No pay for position

    If sponsorships ever exist, they will be visibly marked and will never change the order of recommendations.

  5. Primary sources

    Official specifications, repositories and vendor documentation. We publish no figures without a source.

  6. Public correction

    Vendors and the community can propose changes, which are reviewed before publication.

Inclusion criteria

  • Product available (not just a waitlist) or specification published.
  • Accessible documentation.
  • Fits at least one of the six levels.
  • No practices that contradict the model, such as scraping that ignores robots.txt or bypasses CAPTCHAs.

Scoring weights

For each level, every candidate tool adds up these weighted factors. It is deterministic: the same data gives the same result.

Platform0.25
Level coverage0.35
Maturity0.2
Effort fit0.2

Platform compatibility is a binary filter: a tool that fits neither your platform nor any platform is not a candidate. Tools valid for any platform count as native, so this factor weighs the same for every candidate; integration cost is measured by the effort fit.

Coverage values
Full coverage1
Partial coverage0.5
Maturity values
Stable1
Beta0.6
Draft specification0.3

Deprecated tools are always excluded.

If a tool already recommended for a lower level fully covers this level, it adds 0.1. Only with full coverage. That is why a score can exceed 1.

Diagnostic tools are not candidates: they measure a store's readiness, they do not solve it.

If you prefer stable, draft tools are excluded as well.

Whether a tool is Trusteed's does not enter the score.

Ties are shown as ties and ordered alphabetically by identifier.

Size and goal do not score in version 1.1.0: they are asked, but they do not change the plan.

Rules version: 1.1.0

Freshness

A profile not verified for more than 60 days shows a warning. The warning does not change the order of recommendations.

Corrections

If you spot a wrong figure, a missing tool, or you are a vendor with a right of reply, write to us.

Suggest a correction
Agentic commerce advisor methodology