This page explains how DeciderStack evaluates software, so you can judge how much weight to put on a rating.
How we source a review
We work from what a vendor publishes: pricing pages, plan limits, documentation and changelogs. We capture those pages ourselves and quote the figures as they stood at capture time. Where we have not run a paid plan or a live workload, the article says so instead of implying otherwise.
Data sources
- Live pricing pages, captured on the review date
- Vendor documentation and changelogs
- Cost per request, per GB and per month, calculated from the published rates
- Our own screenshots of the vendor pages the figures come from
Scoring model
We score on four axes: capability (does it do the job), cost predictability (can you forecast the bill), operational burden (how much babysitting it needs), and support quality. A high score on one axis does not cancel a failure on another — we say which axis is the weak one instead of averaging it away.
Update cadence
Pages are reviewed on a rolling basis. Pricing and packaging changes trigger a re-check; a stale page is worse than no page.
Corrections policy
Corrections are applied to the page itself and noted. We do not silently rewrite a verdict to match a vendor’s preference.