Measurement
Part of How to measure prospecting software performance
Why prospecting software benchmark research needs external evidence
External sources and fixed definitions make prospecting software benchmark research credible. Here is the evidence to gather before you compare results.
What to take away
- On 14 August 2025 the Office for National Statistics put professional, scientific and technical services at 13.3% of total UK gross value added, so any lead-generation benchmark has to explain how it maps to that wider economy.
- Internal account data alone flatters prospecting software performance because it ignores the accounts your team never reached.
- Every benchmark needs a written definition of the population, the period and the conversion event before numbers are compared.
- External sources such as the Advertising Standards Authority rulings database and LinkedIn audience data give you market context and precedent that a tool dashboard cannot.
- A benchmark is only useful if a named owner can reproduce it next quarter.
Why is a benchmark not a report?
A report describes what happened. A benchmark sets an expected range so a result can be judged as good, poor or normal. Prospecting software dashboards usually deliver the first and imply the second.
That gap matters at budget review. A reply rate of 4% means little without the rate for comparable teams and sectors. For a full method, see the measurement and reporting guide.
Why does a single tool's data mislead?
Your own platform only records contacts it processed. Accounts never loaded, or rejected at import, vanish from the denominator. Benchmarks built on that base are optimistic by design.
The Office for National Statistics publishes gross value added by industry, which helps you size the sectors you target rather than the sectors you happen to have won. Use it to weight a benchmark, not to replace your own conversion data.
Where should external evidence come from?
Three types of source do most of the work.
Three external evidence sources
Official statistics
- What it provides
- Market size, sector weight
- Example source
- ONS gross value added
- Used for
- Weighting a benchmark
Regulator precedent
- What it provides
- Claims and targeting rules
- Example source
- ASA rulings database
- Used for
- Outbound copy claims
Platform audience data
- What it provides
- Reachable population estimates
- Example source
- LinkedIn audience data
- Used for
- Named job function, region
The Advertising Standards Authority maintains a rulings database covering marketing claims and targeting, which is useful when a benchmark figure is quoted in outbound copy. LinkedIn publishes audience data for B2B targeting that can support a reachable-population estimate for a named job function and region.
How should you define a benchmark?
Write five things before you collect a single number:
- Population
- which accounts, sectors and job functions count.
- Period
- the exact start and end dates, with no rolling window.
- Event
- what counts as a reply, a meeting or a qualified opportunity.
- Exclusions
- internal accounts, duplicates, competitors and test records.
- Owner
- the person who will rerun the calculation.
Ambiguity in any of these turns a benchmark into an anecdote. Attribution choices cause most of the damage, which is why prospecting software attribution methods deserve separate scrutiny.
What does a defensible benchmark look like?
It states the source, the date and the sample, names the records included and excluded, and shows a range rather than one figure.
For example, a team paying £400 a month for prospecting software might report a 6% reply rate on 2,000 contacts. Without the population and period, that number cannot be compared to anything.
Benchmark evidence checklist
- Population defined in writing, with sector codes where relevant.
- Period fixed, with dates recorded.
- Conversion event defined once, and used consistently.
- Exclusions listed and applied automatically.
- Source, date and sample size recorded for every external figure.
- A range published alongside any single figure.
- A named owner and a review date.
- A note on what would invalidate the comparison.
Common questions
How many sources does a benchmark need?
At least one internal dataset and one external source. Two external sources are better when the sectors differ, because they let you show a range rather than a point estimate.
Can a vendor's own benchmark be used?
Only if the vendor states the sample, the period and the method. Check any claim against the Advertising Standards Authority rulings database before repeating it.
How often should benchmarks be refreshed?
Quarterly is workable for most teams, with a full rebuild when the definition of a qualified opportunity changes.
What makes a benchmark fail review?
Undefined populations, mixed periods and figures with no source. If a reviewer cannot reproduce the number from the written method, it will not survive a budget challenge.
