How the benchmarking data is sourced.
A buyer side benchmark is a number a buyer uses to test whether the seller's quote is in band or out of band. The number is only useful if the buyer trusts the source. A benchmark built from analyst reports has the limits of analyst reports. A benchmark built from a vendor's own pricing pages has the limits of vendor marketing. A benchmark built from a single transaction has the limits of a sample of one. Our benchmarking data is built from a specific set of sources, normalised in a specific way, and refreshed on a defined cadence. This page documents each.
The honest version is that benchmark data on enterprise software contracts is hard to get and easy to fake. We have designed the methodology to be transparent about what is real, what is modelled, and what is unknown. Where a benchmark is modelled rather than directly observed, we mark it as modelled. Where the sample is small, we publish the sample size. Where we do not have a defensible benchmark, we say so.
The four sources
The benchmarking data on this site is built from four sources, in order of contribution. Closed engagements where the buyer has consented to anonymised inclusion. Active engagements where the seller's opening position is observable in writing and the buyer has consented to inclusion of the opening position. Public filings, including audited financial statements where vendor spend is disclosed at sufficient granularity. And published academic and industry research where the methodology is documented and the source is verifiable.
The largest contribution by far is the first source. Closed engagements where the executed contract and the prior contract are both visible to the Desk produce the most reliable benchmarks because the comparison is point to point against a verified document. Active engagements contribute opening position data but not closing data. Public filings contribute company level aggregate spend, which we use to triangulate against the engagement data rather than to set benchmarks directly. Academic and industry research contributes context but rarely current pricing.
The normalisation rules
Benchmarks across enterprise buyers are useful only when normalised for the variables that drive the price. We normalise on four dimensions. Deal size, expressed as annual contract value at signature. Region, expressed as the buyer's primary contracting region. Industry, expressed using a published sector classification. And product line, expressed at the level the Broadcom desk negotiates at, which is a level below the product family.
"The benchmark that is not normalised is a number the buyer cannot use. The benchmark that is normalised on four dimensions is a number the buyer can take into a negotiation and defend against the seller."Benchmarking Lead, The Desk
The four dimensional normalisation produces benchmark bands that look narrower than industry analyst figures. The narrower bands are the point. A buyer in financial services in EMEA on a $20M VCF renewal has very little use for a benchmark that averages across all industries, all regions, and deal sizes from $1M to $200M. The buyer needs a benchmark for buyers like the buyer. The normalisation produces it.
Refresh cadence
Benchmark data is refreshed quarterly. The refresh closes the prior quarter's engagement data into the dataset, removes engagements older than 24 months from the active benchmark, and recomputes the bands on the four normalisation dimensions. The refresh date is published with each benchmark figure. A benchmark with a refresh date older than 90 days is flagged as stale on the page where it appears. Stale benchmarks are not removed. They are flagged so the buyer can read the figure with the right context.
What we do not benchmark
We do not benchmark Broadcom commercial terms below a $5M annual contract value. The sample at the lower end is too small for reliable bands and the dynamics differ enough from the enterprise end that the benchmarks would mislead. We do not benchmark commercial terms in regions where we have fewer than five closed engagements in the last 24 months. We do not benchmark indexation language, audit cooperation language, or other contractual terms outside the headline commercial position. Contract language varies in ways that resist normalisation.
We also do not benchmark professional services pricing, support uplift pricing where the support is bundled rather than separately priced, or migration cost estimates from alternative vendors. Each of these is too case specific to produce a useful benchmark band. Where a published benchmark would not be defensible, we tell the buyer that we do not benchmark the item and produce a case specific estimate instead.
Sample size and band publication
Every benchmark figure on the site is published with its sample size and its band. The sample size is the number of engagements that contributed to the figure. The band is the 25th to 75th percentile range across the contributing engagements. The figure itself is the median. We publish the median rather than the mean because the median is less sensitive to outlier engagements. Where a sample size is below 10, the figure is published with a "low sample" flag so the buyer can read it with appropriate caution.
What buyers should use the benchmarks for
Benchmarks are most useful as a buyer side test of whether the seller's opening position sits inside or outside the band where comparable buyers have closed. They are not a substitute for an entitlement reconciliation, a costed alternative, or a structured negotiation. A buyer who walks into a Broadcom renewal with a benchmark and no other preparation will have a slightly better quote than a buyer with no benchmark, and a substantially worse quote than a buyer with a full negotiation preparation. The benchmark is one component of the buyer side toolkit. It is not the toolkit.
Where the benchmarks appear on this site
The benchmark figures appear in three places. The Numbers panel on each product page shows the relevant headline benchmarks for that product. The benchmark articles in each insights category show the detailed bands on specific commercial terms. The calculators in the tools section use the benchmark data as the underlying reference. Every benchmark figure on the site is sourced according to the rules on this page. If a buyer sees a figure on the site that does not appear to be sourced this way, the buyer should write to us and we will correct it.