How we research and pick

Identify what was actually checked

Reading documentation, following a public demo and testing an account establish different things. Each article should identify the work performed and its date. A screenshot of a vendor’s guided tour is labelled as a vendor demo; it does not prove that an import, email or automation succeeded in our own account.

Our small-business CRM comparison uses vendor documentation and US pricing checked on 16 October 2026, an inspected public Close demo, and original buyer worksheets. It does not claim four-product account testing or measured usability.

Build a comparison around the buyer’s job

1. Define the decision. Identify the buyer, market and required workflow before choosing a shortlist. A tool outside a bounded comparison has not been shown to be inferior.
2. Check primary sources. Use the vendor’s pricing, plan, support and export documentation for what it offers. A vendor’s superiority claim is not independent proof.
3. Price the usable plan. Separate annual equivalents from monthly payments, identify feature gates and required fees, and state unverified seat minimums or promotional terms.
4. Separate facts from judgement. A documented feature can support a conditional trial recommendation. It cannot establish an ease-of-use, speed or reliability winner without suitable direct work.
5. Provide something the reader can use. Calculations, comparison criteria and a repeatable trial checklist should help a buyer verify the choice in their own situation.
6. Check the finished article. Reconcile recommendations, prices, links and captions against the supporting evidence. Editing checks do not turn a researched guide into a product test.

When we describe account testing

A hands-on claim needs a record of the account and plan, date, task, observed result and important limitations. A support-response claim needs the actual question, channel and response. Watching someone else’s demonstration or reading their review does not supply those records.

Usability, performance and migration outcomes stay unmeasured unless the article documents direct work. An available export is not proof that a different system can restore every relationship, attachment or conversation.

Shortlists and scores

A shortlist explains who each product fits, the required plan, the reason to skip it and what to verify next. It need not have a universal winner or a numerical score.

Some older review pages carry provisional scores. These are identified as provisional and should not be read as measured performance or verified test results. We withhold Review rating markup while those scores are provisional. The researched CRM guide is unscored.

Scores on this page are provisional until our staff testing round is complete.How we score

Where supported scoring is introduced, its criteria, method and evidence should be visible. We do not collect reader ratings or present editorial scores as reader consensus.

Sources, images and updates

Original account screenshots, vendor-demo captures and conceptual illustrations must be distinguishable. Images support the explanation; they do not establish testing on their own.

Exact prices carry a market, billing basis and access date. We update a review date when substantive checking or revision occurs. A date change alone does not make old evidence current. Check current terms with the vendor before buying.

Affiliate links

Some vendor links may earn a commission if you sign up, at no extra cost to you. A commercial relationship must not determine a recommendation, order or caveat. The article includes its disclosure, and our full affiliate disclosure explains the business model.

Corrections

If a price has changed, a source does not support a claim or an evidence label is unclear, contact the desk. A consequential correction should update the affected recommendations, calculations and review date as well as the sentence itself.