How we work
Methodology
Every number we publish carries a label saying where it came from and how much evidence stands behind it. This page explains what those labels mean.
Real data
Our findings start with people who make enterprise technology decisions: CIOs, CTOs, IT directors and the leaders who sign off on large purchases in the US, UK and Europe. We verify each respondent's role and organisation before including their answers, and we exclude anyone employed by a technology vendor or services firm from buyer samples.
Every survey statistic is published with its sample size and, where relevant, its margin of error. Interview findings say how many conversations they rest on. We take no vendor sponsorship of findings, and nobody outside 22decisions sees results before publication.
Synthetic panels
Some questions are too narrow, too early or too sensitive to field to hundreds of leaders. For those we use synthetic panels: simulated respondent groups built from, and calibrated against, our real survey and interview data.
Synthetic panels help us decide what to ask and where to look. They are not a substitute for buyers. A figure from a synthetic panel is always labelled Modelled, is never used as a headline statistic, and is only promoted to a finding once real respondents confirm it.
Evidence labels
Each statistic on this site and in our reports carries one of four labels, alongside a plain description of its sample, such as “Survey of 412 CIOs and IT directors, Q4 2025”.
- Measured, evidence multiplier ×1
- Taken directly from verified respondents in a single survey or interview programme. The sample size is always stated.
- Corroborated, evidence multiplier ×2
- A survey result that a second, independent method confirms, usually follow-up interviews with a separate set of leaders.
- Tracked, evidence multiplier ×3
- Seen consistently across repeated waves of the 22decisions Buyer Panel and confirmed by a second method.
- Modelled, evidence multiplier ×0
- Produced by a synthetic panel. Useful for direction, never presented as a finding on its own.
The evidence multiplier
The multiplier is a quick way to see how many independent lines of real evidence support a figure. A ×1 figure rests on one source; ×2 on two independent methods that agree; ×3 on two methods and a trend that holds across repeated panel waves. Modelled figures score ×0 because no real respondent has yet confirmed them.
A higher multiplier does not make a number larger or more dramatic. It tells you how safely you can rely on it, whether you are putting it in front of a board or into a sales deck.
Corrections
If we find an error, we correct the report, note the change and date it. To question a figure, write to us through the contact page.