Methodology — How We Source, Verify & Correct Our Data

1. Where the data comes from

Every rate, fee, and percentage on this site traces back to a primary source— the vendor's own public pricing page, official documentation, or published rate card. We do not copy figures from other blogs, aggregators, or comparison sites, because those go stale and propagate each other's errors. Each calculator and guide links to the underlying vendor sources it relies on; if we cannot cite a number to a primary source, we do not publish it.

2. How we verify it — every week

Vendor pricing pages change quietly and without announcements. To catch that, we maintain a data-freshness registry and run a weekly verification pass against every source we track. Software helps us monitor the pages and flag changes, but a human reviews each flagged change and confirms it against the source before any number updates on the site. The most recent verification date is printed on every tool page, so you always know how fresh the figure in front of you is.

When a price does change, we record it in our changelog rather than silently editing the number — so the history is auditable.

3. What our calculators model — and what they don't

The value of an honest calculator is in the costs the vendor's own estimator leaves out. By default we model the edge cases that actually move your bill: currency-conversion markups, international-card surcharges, regional rate differences, fixed-fee drag on small transactions, dispute and chargeback costs, app and per-seat add-ons, and minimums.

We are equally explicit about what a given tool does not yet model — for example, a token-cost calculator focused on text workloads will say plainly that it does not price vision or audio input. Stating the boundary is part of the methodology: a number with an unstated assumption is worse than no number.

4. How accurate the estimates are

Where we can, we validate calculator output against real invoices and billing statements. For the workloads we've cross-checked, estimates land within a few percent of actual billing. The main residual source of error is honest and disclosed — for example, token counts vary slightly between vendor tokenizers for the same input, so a token-cost estimate carries a small inherent margin. We would rather tell you the margin than imply a false precision.

5. Corrections

We will be wrong sometimes — a vendor changes a rate between our weekly checks, or a regional edge case is mismodeled. When that happens we want to know. Report a discrepancy through our contact page and we typically verify and correct within 48 hours, and log material changes in the changelog. A published correction is a feature of a trustworthy source, not an embarrassment.

6. Editorial independence

SmartCloudSuites is supported by advertising and clearly-marked affiliate links — never by influence over our analysis. Where a tool compares vendors, partnerships never change the ranking or the math; the cheapest option is presented as the cheapest option regardless of whether we earn on it. Affiliate relationships are disclosed on our affiliate disclosurepage, and ads never appear inside a calculator's results. Our reputation for neutral math is the only asset that compounds, so we protect it over any single referral.

7. On our use of software and AI tools

We use software — including AI tooling — to help monitor pricing pages, surface changes, and accelerate research. We are transparent about that. What that tooling does not do is decide what we publish: the methodology, the modeling choices, the comparisons, and every figure that reaches the page are reviewed and verified by a human editor against the primary source first. Tools make us faster; they do not replace the verification step.

Questions about how a specific number is derived? Ask via our contact page — methodology questions are the ones we most like to answer. See also our about page for who runs the site and why.