Errors are not silently overwritten
Every report is a versioned research artefact. Anyone quoting a figure from FW Delta Research has to be able to trace later whether it has changed, why it has changed and whether the key finding was affected.
This page describes the version policy, the correction log and the reporting route. It also carries the current version status of every published issue.
Version policy
The version number states how far an issue has changed compared with the version that was quoted.
First publication
The first public version of an issue. Sample, methodology, weights, data cutoff and sources are published in full.
Patch
Typo, broken link or rendering error. The finding does not change. The entry appears in the correction log, date_modified is updated.
Minor
New data, a new source, a changed calculation or a material clarification without a new base methodology. Dependent values, charts and CSV files are regenerated together.
Major
New sample, new weighting or new methodology. The issue is no longer directly comparable with the previous version.
Companion data carry the same version as the report. Older data and document states are archived, not overwritten. Where a figure changes, table, chart, CSV file and claim ledger are updated within the same version.
Correction log
Every material correction appears here and in the version history at the end of the report concerned. An entry always consists of the same fields:
- Date
- Version before and after
- Changed passage
- Reason
- Effect on the key finding
No correction is documented at present
For the published issues, no confirmed correction case exists so far. That is a statement about the current state, not a claim of freedom from error. As soon as a notice has been reviewed and confirmed, it appears here with date, version, changed passage, reason and effect on the key finding.
Until then, the version listed below applies to each issue.
What is recorded internally for every correction
- report_id
- version_before
- version_after
- date
- reported_by
- affected_claim_ids
- old_value_or_text
- new_value_or_text
- reason
- source_or_reproduction
- impact_on_conclusions
Current version status
The edition month is the editorial slot in the series. What counts for citations are first publication, version and data cutoff.
| Report | Edition | First published | Last change | Version |
|---|---|---|---|---|
| FDR-2026-01 SaaS Commercial Transparency Index 2026 | January 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
| FDR-2026-02 Vendor Lock-in & Exit Readiness Index 2026 | February 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
| FDR-2026-03 Software Ownership Economics Report 2026 | March 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
| FDR-2026-04 Automation Unit Economics Benchmark 2026 | April 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
| FDR-2026-05 EU Cloud Switching & Exit Readiness Report 2026 | May 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
| FDR-2026-06 Source-Available License & Governance Risk Report 2026 | June 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
| FDR-2026-07 AI-Assisted Software Delivery Evidence Review 2026 | July 2026 | 29 July 2026 | 29 July 2026 | 1.0 |
Report a correction
Factual notices are explicitly welcome, including from the vendors concerned. Vendors are not asked for approval before publication; a substantiated correction notice is reviewed and, where justified, incorporated transparently.
Subject: Correction notice FW Delta Research. Alternatively through the contact form .
What a notice should contain
- the report ID and version, for example FDR-2026-01, version 1.0
- the passage, table or row concerned
- the disputed value or sentence, verbatim
- the source or the calculation that contradicts it
- optionally the access date of the source
What happens next
- Receipt is confirmed once the notice has been reviewed internally.
- A critical incorrect legal, price or security statement is visibly flagged until it is clarified.
- After the review, the outcome and the reasoning are published, even where the notice does not lead to a correction.
- Fixed deadlines are not promised where a source review takes longer.
What is not a correction
A different opinion on the weighting of a criterion is not a data error, as long as methodology and input values are published correctly. Such notices can be added as a methodology comment or as an additional sensitivity. The same applies to a request to drop a documented limit: limits stay in place.