Every number in the IQ Benchmark Index has a source. This page explains what those sources are, how they're weighted, how we handle uncertainty, and exactly how the index will change as Bundle IQ's transaction data grows.
Most benchmarks don't publish their methodology in detail. We do, for one reason: a benchmark you can't interrogate is a benchmark you can't trust. We want every procurement professional, CFO, and policy researcher who uses this index to understand exactly how confident they should be in each number — and where the limitations are.
We also publish it because we're at an early stage. Bundle IQ launched its competitive sourcing platform in Q1 2025. We have 40+ competitive events in our transaction dataset. That's a real number — not fabricated — but it's a small number. We say so clearly, by category, next to every figure that relies on it. As the transaction volume grows, this methodology page will be updated to reflect a larger proprietary data component. The goal is a benchmark that is primarily powered by real Bundle IQ transaction data. We're not there yet. We're building toward it.
For each benchmark figure, we use the following source hierarchy. Higher-tier sources are preferred where available:
| Tier | Source type | Weight | Current availability |
|---|---|---|---|
| Tier 1 | Bundle IQ competitive event transaction data | Primary (where n≥15) | 40+ events total. Per-category n stated in the index. |
| Tier 2 | Government / regulatory published data | Primary where Tier 1 insufficient | Ofgem tariff data, ONS business expenditure, ABI insurance statistics. |
| Tier 3 | Professional body surveys | Secondary, used to cross-check | CIPS/Hays Salary Survey, REC Fee Survey, BIFM Benchmark, PRCA Fee Survey, Law Society data. |
| Tier 4 | Industry association and trade data | Supplementary where gaps exist | BCS IT data, Logistics UK Quarterly Report, BIBA broker statistics. |
| Tier 5 | Expert analysis | Context and interpretation only — never used as primary rate data | Bundle IQ Research Team practitioner analysis. |
When a competitive procurement event is completed through Bundle IQ, the following data points are recorded anonymously in our category_benchmarks database table:
No supplier names, buyer names, or contract values are stored in the benchmark dataset. The saving percentage and rate range are the only transaction-specific outputs.
We require a minimum of 15 completed events per category before Bundle IQ transaction data is used as a Tier 1 (primary) source. Below 15 events, transaction data is used as supplementary context alongside Tier 2 and Tier 3 public sources. The current n for each category is stated explicitly in the index.
Market rate ranges are calculated as the 20th–80th percentile of observed rates within each category and service type. We exclude the top and bottom 10% of observations to reduce the effect of outliers. The "overpay signal" threshold is set at the 80th percentile — rates above this level are materially above what a competitive process would be expected to achieve.
Where transaction data is insufficient (below the 15-event threshold), rate ranges are derived from the midpoint of available public-source data, with an indicative range of ±15% to reflect the uncertainty in public-source benchmarks applied to specific organisational contexts.
This is not a statistically representative sample of UK SME procurement pricing. The 40+ events in our baseline dataset are not randomly selected — they represent the organisations that have used Bundle IQ for competitive sourcing, which is not a random sample of UK businesses. As the platform scales and transaction volume grows, the representativeness will improve, but it will never be a perfect random sample.
The rate ranges in this index are directionally accurate benchmarks — useful for identifying whether you are materially above market rate — not precise statistical estimates with confidence intervals derived from fully representative samples. We expect future editions, with larger transaction datasets, to allow more precise statements. The Baseline Edition does not.
We believe that a directionally accurate, transparently sourced benchmark is more useful than no benchmark at all, and more honest than a benchmark that overstates its precision. If you use these numbers, use them as they are intended: as a signal of whether a competitive process is worth running, not as a definitive statement of what you should pay.
The most direct way to improve this index is to submit your current rates. The data submission form takes approximately three minutes. All submissions are anonymised before being incorporated into the dataset. You will never be identifiable from the data you submit.
We also welcome academic collaboration. If you are a researcher working on SME procurement, supply chain management, or related fields and would like access to the Bundle IQ anonymised dataset for academic purposes, please contact research@bundleiq.co.uk.