Speed-to-Lead: How to Measure Response Time Properly
It is easy to believe you respond quickly without ever having a number. Measured properly, response time is one timestamp subtracted from another, summarised so that a single average cannot mislead.
By Ben Abel, founder of Ophiro
Published · Fact-checked · 5 min read
The short answer
How do you measure speed-to-lead?
Record the time each enquiry arrived and the time of the first real attempt by a person to contact it, and subtract. Report the median, not the average, together with the percentage of leads attempted within five minutes, within an hour and within a day. Show in-hours and out-of-hours enquiries separately, and do not count automatic acknowledgements as a response.
Key takeaways
- Speed-to-lead = time of first real contact attempt minus time the enquiry arrived.
- Automatic emails and texts do not count. Measure the first attempt by a person or a system that can hold a conversation.
- Use the median and the share within set windows. A mean is dragged up by a few forgotten leads and hides typical performance.
- Measure by clock time and split by when the enquiry arrived. Business-hours-only timing flatters the result.
What is being measured
Speed-to-lead, or lead response time, is how long a new enquiry (inquiry) waits before your business does something about it.
Response time = Time of first real contact attempt − Time the enquiry arrived
Each half of that needs a rule.
The start: when the enquiry arrived
Use the moment the customer acted: the form submission time, the missed-call time or the first chat message. Do not use the time someone opened the notification, or the time the lead was created in the CRM (the system that holds your lead and customer records) if that happens later.
The end: the first real attempt
Three events get confused. Pick one as your main measure and say which.
| Event | Definition | Use it as |
|---|---|---|
| Automatic acknowledgement | A templated email or text sent by software | Not a response. Exclude it. |
| First attempt | A person calls, or sends a personal message that invites a reply | The main measure. It is fully within your control. |
| First conversation | You actually speak or exchange messages | A secondary measure. It depends partly on the customer. |
If an automated system does more than acknowledge, for example by asking and answering questions and booking an appointment, treat it as a real attempt and record it separately from human attempts so the two can be compared.
Why the average misleads
A response time cannot be less than zero and has no upper limit, so one lead forgotten overnight sits far from the rest. A few leads like that dominate an average (the mean).
Illustrative example
Eight enquiries in one week
Response times in minutes, sorted: 3, 6, 9, 14, 22, 45, 180, 960.
- Mean: 1,239 ÷ 8 = 154.9 minutes, about two and a half hours.
- Median: midway between the fourth and fifth values, (14 + 22) ÷ 2 = 18 minutes.
- Within 5 minutes: 1 of 8 = 12.5%
- Within 15 minutes: 4 of 8 = 50%
- Within 60 minutes: 6 of 8 = 75%
- Over 24 hours: 0 of 8. The slowest took 16 hours.
The figures in this example are assumptions chosen to show the arithmetic. They are not benchmarks, do not describe any country’s market and are not drawn from HTJ research. Amounts are in USD for illustration; the sums work the same way in any currency.
The mean says the business takes two and a half hours, while the median says a typical lead waits 18 minutes. Neither is the whole truth. What you want is the median together with the distribution: half within a quarter of an hour, a quarter left for three hours or more.
The measures to report
- Median response time. The middle value once the times are sorted, so half of leads waited less and half waited longer. It is the typical experience.
- Share within fixed windows. This guide uses three: within 5 minutes, within 60 minutes and within 24 hours. Add a window where two of them are far apart and many leads fall in between, as the example above does with 15 minutes. Choose the windows once and keep them.
- Share never attempted. Leads with no recorded attempt at all. Report this separately, and keep those leads in the other three measures as the longest waits of all: they count as a miss in every window and sit at the slow end when you find the median and the 90th percentile. Drop them and the result will flatter you.
- 90th percentile. The time within which nine in ten leads were attempted. It shows how bad the bad cases are.
Clock time or business hours?
Measure clock time: the actual minutes the customer waited. An enquiry sent at 7pm and called at 9am waited 14 hours, whatever your opening hours.
If you open at 9am, counting only business hours turns that into zero minutes and makes the report useless for the question that matters, which is how long customers wait.
To be fair to the team, split the report instead of adjusting the clock:
| Enquiry arrived | Median response | Within 5 min | Within 60 min |
|---|---|---|---|
| In opening hours | |||
| Evenings | |||
| Weekends |
The split shows whether the long waits come from opening hours or from evenings and weekends, and those two problems need different fixes.
Getting the data
You need two timestamps per lead, recorded automatically. The arrival time comes from the form tool, call-tracking system or chat platform. The first attempt comes from the phone system's call log, or from the CRM if calls, texts and emails are logged there as they are made.
Timestamps typed in by hand afterwards are unreliable. If attempts are not being logged automatically, fix that first.
Check before you calculate:
- Time zones. The form tool, phone system and CRM may each be set to a different one.
- Duplicates. One person who submits twice should be one lead, timed from the first submission.
- Test and spam entries. Remove them.
- Inbound calls that were answered. Their response time is zero; include them, or measure forms and calls separately.
Cut the data to find the cause
A single figure tells you there is a problem, and splitting it into segments shows where the problem is.
- By hour and day of arrival, to show gaps in cover.
- By person or team, to show who needs help or capacity.
- By lead source, in case one source notifies the team less reliably than another.
- By channel. Forms, missed calls and chats may follow different routes inside the business.
Connect speed to outcomes
The reason to measure response time is its link to results. Once you have both timestamps, group leads into bands by response time and compare each band's contact rate and qualification rate. For the qualified column, use a written definition applied the same way to every lead, as set out in how to calculate cost per qualified lead.
| Response time band | Leads | Reached | Qualified |
|---|---|---|---|
| Under 5 minutes | |||
| 5 to 60 minutes | |||
| 1 to 24 hours | |||
| Over 24 hours |
This gives you your own evidence, from your own market and customers, of how outcomes differ with delay. A published study describes other companies, so this is the better basis for deciding how much to invest in faster cover. Read the bands with care if the numbers in each are small, and remember that leads answered slowly may differ in other ways, such as arriving out of hours.
Common mistakes
Four of these have come up already: reporting the mean alone, counting the automatic email as the response, excluding leads that were never attempted, and measuring business hours only. Two more are easy to make.
- Measuring once. Response time can drift as volume and staffing change. Report it weekly beside lead volume.
- Setting a target nobody can see. Put the median and the within-five-minutes share where the team sees them.
What to aim for
How fast is fast enough is a separate question, covered in how quickly you should respond to a kitchen or bathroom enquiry. Where enquiries carry a fixed deadline, how to qualify a bridging finance enquiry shows routing by the days left. Whatever target you choose, state it in the same terms you measure: a median, and a share within a window.
Limitations
- No benchmark response times are given. HTJ has not yet measured them in any market.
- The method assumes your systems record accurate timestamps. Check time zones and clock settings before trusting the output.
Run the numbers
- High-Ticket Lead Economics Calculator
Model a high-ticket lead funnel from ad spend to gross profit: cost per enquiry, cost per qualified lead, customers, ROAS and break-even cost per click.
Spotted an error? Tell us and we will correct it. Read our editorial policy.

About the author
Publisher, High Ticket Journal. Founder, Ophiro
Ben Abel publishes High Ticket Journal and is the founder of Ophiro, a customer acquisition agency that builds Google Ads, landing page and lead qualification systems for high-ticket businesses.
