Methodology
Step 01
Report ingestion
Submit any publicly accessible URL or upload a PDF. Select your sector for tailored UK regulatory context.
Step 02
Pattern analysis
Examined against nine statistical manipulation patterns and cross-referenced against UK ICO data, DSIT 2025 figures, and NCSC CAF 4.0.
Step 03
IGP-grounded questions
Each finding maps to a CAF 4.0 outcome and generates questions anchored to Indicator of Good Practice criteria for vendor briefings.
Nine patterns detected
| Pattern | What it means | UK example | |
|---|---|---|---|
| 🍒 | Cherry picking | Selective time windows or geographies to manufacture a favourable trend. | Report window starts right after a trough. |
| 📊 | Misleading averages | Mean used where median is more appropriate, distorting the representative figure. | Breach cost inflated by nation-state incidents, presented to UK SMEs. |
| 📉 | Truncated Y-axis | Axis not starting at zero amplifies modest changes into apparent crises. | A 2% YoY increase rendered as a near-vertical spike. |
| 🔬 | Small sample size | UK-wide conclusions from self-selected respondents without statistical adjustment. | "UK market" findings from 180 enterprise respondents on a vendor mailing list. |
| 🔗 | Correlation vs causation | Correlation presented as causation without forensic validation. | "Our product = 60% fewer incidents" without controlling for organisation size. |
| 🔢 | % vs absolute numbers | Percentages obscure small absolute base numbers, or vice versa. | "Ransomware up 100%" masking an increase from 12 to 24 cases. |
| 💰 | Sponsor bias | Findings systematically support the funding organisation's interests. | All recommended controls align with the sponsor's product portfolio. |
| 🏷️ | Definition laundering | Vague terminology makes claims unfalsifiable or broadens scope post-hoc. | Signature-matching rebranded as "AI-powered threat intelligence." |
| 🌍 | Geographic sample mismatch UK-specific | US or global data applied to UK context without adjustment. | IBM CODB £4m+ cited to UK SMEs without noting DSIT 2025 UK mean of £1,970. |
All outputs are probabilistic observations of statistical methodology. Not professional advice or audit findings. Verify independently before acting. NCSC CAF 4.0: August 2025 · ICO: Q4 2025 · DSIT: 2025.