Speed to lead (also called lead response time) measures how quickly someone follows up after a buyer asks to hear from you. It's one of the most actionable metrics in the lead funnel, but a single number such as "19 hours" means little until you say: from when, to what, in which hours, and what happened to the leads nobody contacted.
Worked example
Ten demo requests arrived in a week, with a service-level agreement (SLA) of a first touch within one business hour. Response times in business hours:
0.2 · 0.4 · 0.5 · 0.8 · 1.5 · 3 · 5 · 24 · 60 · never touched
| Measure | Calculation | Result |
|---|---|---|
| Average (touched leads only) | 95.4 ÷ 9 | about 10.6 hours |
| Median, touched leads only | 5th of 9 values | 1.5 hours |
| Median, never-touched counted as slowest | middle of 10 values: (1.5 + 3) ÷ 2 | 2.25 hours |
| Within SLA (under 1 business hour) | 4 ÷ 10 | 40% |
| Tail: 9 out of 10 touched within | 9th value | 60 business hours |
| Never-touched rate | 1 ÷ 10 | 10% |
The average is dragged up by two slow leads and describes nobody. The median describes the typical lead. The SLA percentage shows how often the promise was kept, and the tail shows the worst experience.
Start and stop
Define both events and record them as timestamps:
- Start: the moment the lead became sales-ready, for example when a demo form was submitted or the lead was marked as marketing-qualified. Reassigning the lead doesn't restart the clock: the buyer has still been waiting.
- Stop: the first qualifying touch: a call logged by the dialer (answered or not), or a personal one-to-one email or message from the owner, logged in the CRM.
- Doesn't count: automated or sequence emails sent without review, a status change to "Attempting", a task created, or a call typed in by hand with no dialer record.
Measure from activity records, never from status fields. If changing a status stops the clock, people will change the status. It's the same principle as buyer-verified stage exit criteria: count what happened, not what someone clicked.
Business hours and clock hours
With a 9-to-6 working day, a demo request submitted at 4:00 p.m. on Friday and called at 10:00 a.m. on Monday waited 66 hours on the clock, but only 3 business hours (2 on Friday, 1 on Monday). Report both and label them: business hours are fair to the team and are what most SLAs use; clock hours are what the buyer experiences. Store raw timestamps and calculate business time in reporting, using a business-hours calendar for each region.
Never-touched leads are failures
Leads that were never contacted have no stop time, so many reports quietly drop them. That flatters the result: the worse the follow-up, the more leads drop out, and the better the median can look. Instead:
- Report a never-touched rate: leads with no qualifying touch by your deadline ÷ all leads.
- In SLA compliance, keep never-touched leads in the denominator as failures.
- In the median, treat them as slower than any touched lead, as in the example. A median only needs the values in order, so this works; an average can't do it.
Only measure leads whose SLA window has closed. A lead that arrived ten minutes ago hasn't failed yet.
Common mistakes
- Reporting the average. One very slow lead moves it more than everything else combined.
- Reporting the median alone. It ignores the tail completely. Pair it with % within SLA and a tail measure such as "9 out of 10 leads were contacted within X hours".
- Excluding never-touched leads. It rewards the worst follow-up.
- Stopping the clock on a status change or automated email. Easy to game, and invisible to the buyer.
- Claiming speed causes conversion without checking mix. Fast responses often go to the best leads, such as demo requests. Compare like with like, for example win rate by lead source and response time, before concluding.
The free RevOps Foundations course covers speed-to-lead as part of the marketing-to-sales handoff in Lesson 2, "One revenue lifecycle: the bowtie".