How we measure S1 research
Where every S1 figure on this site comes from: what each term means, which production runs the figures describe and the outside benchmarks we quote.
Updated October 20, 2026. SalesOne research team. Back to S1 agents
What do the terms mean?
The same words are used the same way on every page that quotes these figures.
- Kept
- A company S1 agents accepted against the profile and delivered as an account. Also called accepted or delivered.
- Reject reason
- The rule a company failed and the evidence, stored with the company. No company can be turned down without one.
- Dated
- A buying signal dated to the day the event happened, or to the month when the source gives no day.
- Ready person
- A person with a public source dated within the last 12 months and a profile match. Only ready people go to contact lookup.
- Data sources
- The distinct public sources cited in delivered research: 1,000+ so far, such as SEC filings, press releases, company sites, trade press and public records.
Which production runs are the figures from?
The figures describe what S1 agents delivered in SalesOne’s October 2026 production research runs: every company they checked, every reason, every signal and every person. The reject data is broken down by reason in its own guide.
- Runs
- SalesOne production research runs, October 2026
- Companies checked
- 1,051, with 249 kept and 802 turned down: about three in four turned down, each with its reason
- Rejects
- All 802 keep the reason they failed; 88% of reasons contain a date. Top reason: no qualifying event inside its window (67%)
- Buying signals
- All 375 link their source, and 98% are dated to the day
- People
- About four mapped per account, 95% with a dated source
- Review pass
- A second check that re-reads every buying signal at its source before delivery
- Repeat research
- None: nothing is researched twice for a client
Each figure describes an account as it stood when it was delivered.
Which outside benchmarks are quoted?
WideSearch (arXiv 2508.07999, August 2025) asks agents to collect every entity in a set into a table; the best agent tested fully completed 5.1% of tasks. Dedicated deep research products were left out because they returned reports instead of tables. DeepWideSearch (arXiv 2510.20168, October 2025) adds multi-step reasoning to find each entry; the best average success was 2.4%. Both measure general AI agents; neither tested S1.
See your own market, researched.
Bring your offer and one customer you would like more of. We will research a sample of your market and show you the accounts S1 agents accept, the ones they turn down, and why.