Reviews
Part of A rigorous method for reviewing small business CRM products and services
How CRM products are reviewed here, and when scores are withheld
A disclosed CRM review method for England, covering evidence hierarchy, controlled tests, conflicts, unknowns and rules for withholding scores.
This methodology treats a review as an auditable answer to one buying question. It does not award a universal score to a product whose value depends on edition, configuration, data, users and contract.
What to take away
- No numerical score is issued because value depends on edition, configuration, data, users and contract.
- Supplier pages show what a supplier says, not independent performance, security or accessibility.
- Missing evidence is marked unknown rather than converted into a low score.
- Mandatory gates are independent, so one cannot compensate for another.
- Any change to edition, terms, processing, integration or workflow triggers a review.
Review disclosure
- Reviewer role:
- Relationship Brief editorial evidence reviewer. The named individual remains to be assigned before publication.
- Research date:
- 6 September 2026, with a publication-day recheck required.
- Subject:
- exact small-business CRM editions and CRM implementation services available to an organisation operating in England.
- Method:
- desk review of official, regulator and first-party records, followed by a proposed buyer-controlled test method.
- Hands-on boundary:
- no account, integration, import, support interaction or supplier project was tested.
- Conflicts:
- no supplier funding, referral commission, free account or commercial relationship was disclosed for this review.
- Scoring rule:
- no numerical score is issued. Mandatory gates use pass, fail, unknown or not applicable.
Define the comparison case
Record the workflow, user roles, eligible records, required state changes, evidence outputs, integrations, data geography, contract term and exit need. Add the current spreadsheet or system as a baseline. A candidate is out of scope when its exact edition, operating region or commercial unit cannot be established.
Define the Comparison Case
- Workflow and user roles
- Eligible records and state changes
- Evidence outputs and integrations
- Data geography and contract term
- Exit need and current baseline
- Out of scopeunclear edition or region
This prevents a supplier's broad product family from being compared with another supplier's entry plan. It also separates a feature's existence from its fitness for the buyer's configured process.
Use evidence for what it proves
The hierarchy is buyer-controlled test evidence, applicable official or regulator records, exact supplier documentation, independently reproducible technical material, then customer commentary. Supplier pages can establish what a supplier currently says. They do not establish independent performance, security, accessibility or suitability.
Evidence Hierarchy and Limits
Evidence type
- Buyer-controlled test
- Strongest
- Official or regulator records
- Strong
- Exact supplier documentation
- Moderate
- Reproducible technical material
- Moderate
- Customer commentary
- Weakest
- Supplier pages
- Limited
What it proves
- Buyer-controlled test
- Actual performance
- Official or regulator records
- Compliance facts
- Exact supplier documentation
- Supplier claims
- Reproducible technical material
- Technical behavior
- Customer commentary
- Experiences only
- Supplier pages
- Current statements only
The CMA's online review-site guidance addresses fair checking and interests in published reviews. Its existence does not validate individual CRM reviews, so provenance and incentives are recorded rather than assumed.
For a cloud CRM, the NCSC's service-selection guidance supports defining needs and testing provider evidence against relevant security principles. Security remains unknown until the buyer's qualified reviewer has examined the planned configuration and supporting material.
Screen, then test
The desk screen checks exact edition, region, commercial unit, documented workflow capability, import and export, contract record, privacy materials, security evidence, accessibility evidence and support route. Missing evidence is marked unknown. It is not converted into a low score or filled from a different plan.
A controlled test then uses synthetic contacts, organisations and deals. Each candidate receives the same task script: import, deduplicate, assign, change status, suppress a marketing route, export and recover. Capture the environment, user role, configuration, expected result, actual evidence and defect. Negative paths matter because a polished happy-path demo says little about exception handling.
The W3C's evaluation overview makes clear that accessibility methods have different purposes and no single tool is sufficient. The review therefore records which user journeys, assistive technologies and human checks were actually completed, leaving the gate unknown when they were not.
Decide and expire the record
Privacy and PECR, security, accessibility, functional fit, contract, finance, migration, support and exit are independent gates. One cannot compensate for another. Tradeable preferences may be described after all mandatory gates pass, but a weighted total must disclose its owner, weights and sensitivity.
The possible conclusions are proceed to a controlled trial, seek evidence, negotiate a remedy, retain the baseline or reject. The record names the decision owner, unresolved questions, source dates and expiry. Any edition, terms, processing, integration or business-workflow change triggers review. Qualified legal, privacy, security, accessibility, commercial and finance reviewers must approve their own gates before publication or purchase.
Before you act
- Record the workflow, roles, records and required state changes.
- Confirm the exact edition, region and commercial unit.
- Mark missing evidence as unknown, not as a low score.
- Run the same controlled test script on each candidate.
- Name the decision owner, unresolved questions and expiry date.
- Get qualified reviewers to approve their own gates before purchase.
Common questions
Why does this methodology withhold a numerical score?
It treats a review as an auditable answer to one buying question. A universal score cannot fit a product whose value depends on edition, configuration, data, users and contract. Mandatory gates instead use pass, fail, unknown or not applicable.
What evidence counts most in this hierarchy?
Buyer-controlled test evidence ranks first, followed by applicable official or regulator records, exact supplier documentation, independently reproducible technical material and customer commentary. Supplier pages only establish what a supplier currently says, not independent performance, security, accessibility or suitability.
What happens when evidence is missing?
Missing evidence is marked unknown. It is not converted into a low score or filled from a different plan. The accessibility gate stays unknown when user journeys, assistive technologies and human checks were not completed.
