Our methodology

What we cover and why

We research software categories where the decision is consequential and the marketing is loud. Within each category we select products that represent a realistic buyer's shortlist — not every product that exists. Selection is editorial: vendors cannot pay to be included, request inclusion, or see reviews before publication.

The research process

Every category moves through the same journey — the one our site structure mirrors:

• Insights — we map the market: pricing models, consolidation, the differences that actually change outcomes.
• Guides — we establish decision criteria from real buyer questions before evaluating anyone.
• Compare — products are assessed against those criteria using vendor documentation, published pricing, product demos and trials, and patterns in long-term user feedback.
• Reviews — individual products receive scores and verdicts, including honest weaknesses.

Where a review states hands-on evaluation, it means we ran the product on our own accounts with real data. Where it does not, the assessment draws on documented features, pricing analysis and corroborated user experience — and we say so.

The five-stage decision-support matrix

Behind every review sits an internal evaluation sequence. Each stage can disqualify or reshape what the next stage concludes:

1. Compliance check — does the product meet the regulatory reality of UK buyers? (MTD recognition for tax software, FCA status and FSCS versus safeguarding for banking, UK data-residency options for hosting.) Failures are disclosed prominently, not averaged away.
2. Capability mapping — we map the product against the buyer's actual workflow before scoring anything, drawn from real buyer questions rather than vendor feature lists.
3. Sentiment analysis — patterns in long-term user feedback (not launch-week reviews) are weighted: support quality on the bad days, not the onboarding days.
4. Cost-to-value modelling — we model total cost at realistic volumes: three-year payroll costs by team size, bank fees across five usage meters, hosting at renewal prices rather than intro pricing. This is the stage most comparison sites skip; our calculators exist to show its working.
5. Peer review — scores and verdicts are challenged against the rest of the category before publication, so a 9.0 means the same thing in banking as in hosting.

AI-assisted research, human-verified

ReQompass uses AI tooling in the research and drafting process — and treats it the way a good newsroom treats an intern: fast, broad, and never published unchecked. The rules:

• Every product profile carries a verification record: the date a human last re-checked it against vendor-published sources, linked visibly in the data-integrity strip at the top of each review.
• Profiles awaiting their next human audit say so plainly — we would rather show you an honest pending state than a false verified one.
• Our data-freshness audit flags any verified record older than 90 days and blocks silent ageing of the archive.
• Primary-source links on reviews let you check every consequential figure yourself in one click.

How we score

Products receive a score out of 10, weighted across five criteria. Weights vary by category — security categories weight trust architecture more heavily; collaboration tools weight adoption more heavily:

What the scores mean

• 9.0+ — Recommended to most buyers in the category without reservation.
• 8.0–8.9 — Excellent with stated trade-offs; the right choice for the right buyer.
• 7.0–7.9 — Competent with material weaknesses; buy only for specific needs.
• Below 7.0 — We would typically steer buyers elsewhere. (We rarely publish reviews of products this weak; we tell you in comparisons instead.)

How we handle money

ReQompass is reader-supported. When you buy through links on our site we may earn a commission, at no extra cost to you. Our rules:

• Commission terms are unknown to the person scoring the product and are never a scoring input.
• We review products that pay us nothing alongside those that do, on identical criteria.
• Rankings are never for sale, to anyone, at any price.
• If these rules ever conflict with a commercial arrangement, the arrangement goes.

Updates and corrections

Software changes constantly, so reviews are reassessed on a rolling basis and whenever pricing or products materially change. Every page shows its last-updated date, and every review shows its own verification record in the data-integrity strip — including an "audit due" state when a record passes 90 days. If we get something wrong, we correct it promptly and note significant corrections on the page. Corrections from readers and vendors (with evidence) are welcomed, investigated and acted on regardless of any commercial relationship.

Limitations — stated plainly

We cannot run every product in every environment, and prices shown are indicative entry-level figures that vendors change without notice. Our scores are expert editorial judgment, not laboratory measurements. Always verify current plans on the vendor's site, and trial the product with your real workload before committing — our guides are designed to make that trial efficient.