How to Measure Lead Response Time

Ask a business how fast it replies to leads and you'll usually hear "pretty quickly". This guide turns that into a number you can trust. You need a spreadsheet, about 20 leads and half an hour.

Decide what starts and stops the clock

Most bad figures come from getting these two points wrong.

Start the clock when the enquiry was submitted, rather than when your team noticed it. A lead that sat in an inbox for two hours has waited two hours, whoever saw it. Lead data usually includes this time. Meta's export has a created_time column and Google's lead form webhook sends a submit time.

Stop the clock at the first attempt by a person: a call, or a text or email written to that lead. An automatic reply doesn't count. It's a good habit, but nobody has dealt with the enquiry yet.

If you want a second, harder measure, also record the time of the first real conversation. The first tells you how fast your team moves. The second tells you how fast customers get help.

Collect the data

Set up a sheet with these columns, and put each lead on its own row:

• A: Lead name
• B: Received (date and time)
• C: First attempt (date and time)
• D: Minutes
• E: Source
• F: Owner

Fill in your last 20 to 30 leads, using call logs and sent messages to find the times. Leave First attempt empty if nobody tried. If you can't find the times, start logging from today. Don't rebuild the numbers from memory, because memory is always kinder than the call log.

A worked example

Say a small heating firm had these 11 leads from Monday to Thursday.

Lead	Received	First attempt	Minutes
A	Mon 09:12	Mon 09:20	8
B	Mon 11:40	Mon 12:05	25
C	Mon 15:03	Mon 15:09	6
D	Mon 18:47	Tue 08:30	823
E	Tue 08:05	Tue 08:17	12
F	Tue 10:30	Tue 14:10	220
G	Tue 13:22	Tue 13:40	18
H	Wed 09:55	Wed 10:40	45
I	Wed 12:15	Wed 12:23	8
J	Wed 16:20	Thu 09:05	1,005
K	Thu 07:50	None	None

The four numbers that matter

1. Median response time

Sort the minutes from lowest to highest and take the middle value. With an even number of values, take the halfway point between the middle two.

Sorted, the ten contacted leads are 6, 8, 8, 12, 18, 25, 45, 220, 823 and 1,005. The middle two are 18 and 25, so the median is 21.5 minutes.

The average of those same leads is 217 minutes (2,170 ÷ 10), which is ten times higher. Two slow leads have dragged it up, and it describes nobody's actual experience. The median tells you what a typical lead gets.

2. 90th percentile

A median can hide your worst delays, so you need a second number: how long your slowest leads wait. The 90th percentile is the time that 9 in 10 of your leads beat.

For ten leads, it's the 9th value in the sorted list: 823 minutes, or 13 hours 43 minutes. If your median looks good and this number looks terrible, your process works most of the time and fails badly at certain times.

3. Share contacted within your target

Leads with a first attempt within the target ÷ all leads received

With a 30-minute target, 6 of the 11 leads made it: 55%. Use all 11 as the bottom number, including the lead nobody contacted. This is the easiest number to share with a team, because everyone understands it.

4. Share never contacted

Leads with no attempt ÷ all leads received

Lead K makes it 1 of 11: 9%. Report this separately, every time. The median and percentile ignore K because it has no time to measure, so without this number, your worst result would vanish from the report.

The spreadsheet formulas

These work in both Excel and Google Sheets, using the columns above with the data in rows 2 to 12. Format columns B and C as date and time.

• Minutes (cell D2, copied down): =IF(C2="","",(C2-B2)*1440)
• Median: =MEDIAN(D2:D12)
• 90th percentile: =PERCENTILE(D2:D12,0.9)
• Within 30 minutes: =COUNTIF(D2:D12,"<=30")/COUNTA(B2:B12)
• Never contacted: =COUNTIF(C2:C12,"")/COUNTA(B2:B12)

The percentile formula gives 841 here rather than 823, because spreadsheets estimate between values. Either is fine, as long as you use the same method each time.

Clock time and working hours

The clock figures above treat lead D as a 13-hour failure. But D came in at 18:47, after closing, and got a call 30 minutes after opening the next day. Was that really a problem?

So work out a second set of figures that only counts working hours. Say this firm is open 08:00 to 18:00, Monday to Friday:

• D counts from 08:00 on Tuesday, which makes it 30 minutes.
• J counts from 16:20 to 18:00 on Wednesday, then 08:00 to 09:05 on Thursday, which makes it 165 minutes.
• Everything else stays the same.

Measure	Clock time	Working hours
Median	21.5 minutes	21.5 minutes
90th percentile	823 minutes	165 minutes
Within 30 minutes	55%	64%
Never contacted	9%	9%

The two views answer different questions. Clock time is what the customer experienced, and it's how you judge your out-of-hours plan. Working hours is what your team controls, and it's the fair way to judge them. The second view also moves attention to the right leads. D looks fine now. F (3 hours 40 in the middle of a Tuesday) and J (a mid-afternoon lead left overnight) are the ones to ask about. And K was received ten minutes before opening and still got no call at all.

For a small volume, working this out by hand takes a few minutes a month.

Find the cause by splitting the numbers

One overall number tells you there's a problem. Splitting it tells you where.

• By source. If Facebook leads have a median of three hours and website leads have one of 15 minutes, suspect how Facebook leads reach you, and not the person answering. See how to manage Facebook and Instagram leads.
• By owner. Use it to spot who is overloaded or has the wrong alert settings. Be careful with this one. Someone on site all day isn't comparable with someone at a desk.
• By time of day and day of week. Regular slow patches usually point to a cover problem, such as lunchtimes or the school run.

Check it against your own results

Once you have 100 or more leads with outcomes, group them by response time, for instance under 30 minutes, 30 minutes to 4 hours, and over 4 hours. Then compare the win rate for each group.

Don't read too much into it. Urgent jobs tend to get faster replies and are also more likely to close, so speed may not be the only cause. Still, it's your own evidence, which is far more persuasive to a team than anyone's blog post.

Keep a simple weekly scorecard

Five figures, on one line, each week:

• Leads received
• Median (working hours)
• 90th percentile (working hours)
• Share within your target
• Never contacted

If you get fewer than about ten leads a week, use a rolling four weeks, because a single lead will swing the numbers. Set your target based on where you are now. If your median is 45 minutes, aim for 30 before you aim for 5. To bring these numbers down, work through how to respond to new leads faster.

Mistakes that flatter the numbers

• Using the average alone.
• Leaving out leads nobody contacted.
• Counting automatic replies as a response.
• Starting the clock when someone saw the lead.
• Comparing people without looking at what else they were doing.

Measuring with LeadCentre

LeadCentre doesn't calculate a response time figure for you today, so the sheet above stays useful. What it does is work on the gaps that the figures reveal. The Android app sends a push notification for each new lead, the lead screen has one-tap contact and templates, and unread leads are marked with a lime border and dot so they stand out in the list. When you connect a source by webhook, you can match its submit time to the Date received field, so the record shows when the lead really came in.