Skip to content
Methodology

Data and verification policy

How provider records are built, what each verification label means, how corrections and case studies are handled, and how paid placement is kept separate from ranking and matching.

Where data comes from

Every provider record stores its data source. Records are built from four kinds of source, in this order of preference:

  1. 1
    Public company websites

    Services, capabilities, technologies, locations, team size and published pricing pages.

  2. 2
    Partner directories

    Cloud, model-provider and platform partner listings, used to corroborate technology experience.

  3. 3
    Public case studies

    Case studies, talks and press published by the provider or its client, linked from our record.

  4. 4
    Provider submissions

    Profiles submitted or claimed by providers, reviewed by the platform team before going live.

Fictional sample records used for development are labelled Sample, shown only in demo mode and excluded from all public statistics.

The three verification tiers

Every provider profile and case study carries exactly one of these labels. The label describes how the record was checked, not how good the provider is.

Unverified

What it means
The profile was created by the platform team from public information (website, partner directories, public case studies) and has not yet been checked by a second reviewer or confirmed by the provider.
What it does not mean
It does not mean the provider is unreliable. It means treat the details as sourced from public material that may be out of date.
How it is set
Automatic on creation.

Platform verified

What it means
A member of the platform team checked the profile's core facts (company exists, services match the listed capabilities, location, team size band, published pricing) against at least one public source, and recorded the source and date.
What it does not mean
It is not an assessment of delivery quality, and the provider may not yet have seen the profile.
How it is set
Platform review. The verification source and date are stored with the record.

Provider verified

What it means
The provider claimed the profile with a verified company email address and confirmed or corrected its details. Provider-verified profiles are the provider's own statement of what they do.
What it does not mean
It is not an endorsement, and claims within the profile are still the provider's, not ours.
How it is set
Claim flow: company-domain email verification, then confirmation from the provider dashboard.

Claimed vs platform-created profiles

A platform-created profile was researched and written by our team from public sources. The provider may not know it exists. It is labelled “Platform-created profile” and starts as Unverified or Platform verified.

A claimed profile has been taken over by someone at the provider who verified a company-domain email address. From that point the provider controls the description, capabilities, pricing and case studies, and the profile is labelled “Provider claimed this profile”. Claiming plus confirmation results in the Provider verified label.

Claiming is free. Nothing about claiming affects ranking or matching.

How corrections work

  1. Anyone can request a correction: providers through the dashboard or the contact form, buyers and third parties through the contact form, naming the record and the correction.
  2. The platform team checks the correction against a public source or evidence supplied by the provider (for example a contract-free statement of services, a partner listing, or a public case study).
  3. The record is updated and the review date is refreshed. Material changes to a Platform verified record trigger a re-check of the affected fields.
  4. Disputed facts that cannot be settled from public sources are removed rather than guessed.

We aim to act on corrections within five working days. Request a correction.

How case studies are sourced and linked

Case studies are the evidence layer of the directory, so they follow stricter rules than profiles:

  • Each case study records the problem, solution, results, team size, duration, stage reached (for example pilot to production), integrations, technologies and capabilities used.
  • Where a public source exists (the provider’s or client’s own publication, a talk, a press article) the record links to it with a source label. Readers can check the original.
  • Case studies without a public source are displayed with the note “No public source on file — treat as provider-reported”.
  • Case studies carry the same verification labels as profiles and are counted as matching evidence only for relevance to a deployment type or industry, never as a quality score.
  • Client names are shown only when the source discloses them. Otherwise the client is described by industry and size.

How featured and sponsored placement is labelled

Providers and sponsors can pay for visibility. To keep the directory useful, paid visibility is separated from the parts of the platform that buyers rely on for judgement:

  • Labelled everywhere. Featured providers carry a Featured badge on every card and profile. Sponsored articles, case studies and research carry a Sponsored badge and a disclosure naming the sponsor.
  • Excluded from ranking. Directory sort orders (relevance, name, newest, completeness) ignore featured status. Featured slots are rendered as a separate, labelled block.
  • Excluded from matching. The matching engine does not read featured or subscription status. A sponsored match, where offered, is shown as an additional labelled result, never mixed into the scored list.
  • No ratings. We publish no star ratings, “best of” lists or awards, paid or otherwise.

How matching weights are published

Matching produces a score from 0 to 100 for each provider against a deployment profile. The score is a weighted sum of eight factors. The default weights are below; if the platform team adjusts them, the current values are shown on the match page under “How matching works”. Every provider is scored with the same weights.

Default matching weights
FactorWeight
Capability fit30%
Industry experience15%
Technology experience15%
Engagement model10%
Geography and timezone10%
Team size10%
Budget fit5%
Availability5%

Each result explains its reasons and gaps in plain language. See how matching works.

Data retention for leads

When a buyer submits a requirement, contacts a provider or requests a resource, we store the submission as a lead so that introductions can be followed up and traced.

  • Lead records are retained for 24 months from the last activity, then deleted or anonymised for aggregate statistics.
  • A lead is shared with a provider only when the buyer contacts that provider directly or agrees to be introduced to matched providers.
  • Buyers can ask for their lead to be deleted at any time through the contact form; providers receiving an introduction are notified of the deletion.
  • Assessment and calculator inputs submitted without contact details are stored anonymously.

Full details are in the privacy policy.

Frequently asked

Does paying for a featured listing change a provider's verification status?
No. Verification is earned only through platform review or provider confirmation. Paid placement is labelled and has no effect on verification, ranking or matching.
How do I correct information about my company?
Claim the profile and edit it, or send the record and the correction through the contact form. Corrections are checked against public sources or provider evidence and applied with a note of what changed.
Are the matching weights public?
Yes. The default weights are capability 30%, industry 15%, technology 15%, engagement model 10%, geography 10%, team size 10%, budget 5% and availability 5%. They are shown on the match page and are the same for every provider.
How long do you keep buyer requirements?
Lead records are kept for 24 months from last activity so that follow-ups and introductions can be traced, then deleted or anonymised. You can ask for deletion earlier at any time.
Are the sample providers real?
No. Records labelled Sample are fictional and exist only to exercise the platform in demo mode. They are excluded from public statistics.