Sales development generates a great deal of published research and not much of it is independent. That is not disqualifying. It does mean reading it takes a method.
Four checks before you use a number
1. Who paid for it? Vendor-sponsored research is not fraudulent, but the questions were chosen by someone with an interest in the answers. A study finding that teams using category X outperform those that do not, published by the leading vendor in category X, is a marketing artefact with a methodology section.
2. Who answered? Sample size is the number everyone quotes and composition is the one that matters. Two hundred responses skewed to one segment, one region, and one company size does not describe your situation. Look for the breakdown; its absence is itself informative.
3. What was actually asked? Self-reported metrics are notoriously unreliable in this field, because definitions vary so much. A survey asking for connect rate gets answers counted at least two different ways, which makes the average close to meaningless. See the glossary for how contested these terms are.
4. Is the distribution shown, or only the average? An average connect rate across a bimodal population describes nobody. Research that publishes quartiles is more useful than research that publishes a mean, and far rarer.
Where the useful benchmarks come from
In rough order of reliability for a given question:
- Your own historical data. Always the best benchmark for you, and the one
teams skip because it takes two quarters to build.
- Aggregated operational data, where a provider measures what actually
happened rather than asking people to recall it.
- Independent surveys with published methodology.
- Vendor-sponsored surveys, read with the four checks above.
- Conference-stage numbers, which should be treated as anecdote unless
sourced.
What Tenbound publishes
The research library holds the formal output with figures and references. The resource library holds the wider archive, and playbooks hold the procedures rather than the findings.
Everything is free to read and free to cite. The one ask is attribution: name Tenbound and link the source. That principle is the reason the research exists in the first place, and it is what keeps the institute distinct from the commercial brands in the group.
On benchmarks you cannot verify
A note that applies to this whole category, including some of Tenbound's own older material: where a benchmark's underlying data is not available to re-examine, it should be read as directional rather than as a target. Several pages in this archive previously carried survey findings whose data cannot now be reproduced, and those pages state that rather than restating numbers that cannot be checked.
Where this sits
Research is the foundation the Tenbound Pipeline Architecture Standard is built on, and Measurement is the pillar that consumes it. A benchmark applied without checking its definition against yours is the most common way a team concludes it is underperforming when it is measuring differently.