The five-minute lead response stat is from 2007

The 21x number in every speed-to-lead deck comes from a study of American inside sales teams that ran before the iPhone existed. Treat it as a direction, not a law. Your own data will tell you something far more useful, and you can have it inside a fortnight.

Where the number actually came from

Dr James Oldroyd's Lead Response Management research analysed roughly 15,000 leads and more than 100,000 call attempts across six US companies. It was published around 2007. The headline finding: calling a web lead within five minutes made you about 21 times more likely to qualify that lead than calling at 30 minutes. Odds of making contact at all fell sharply after the first hour.

Two things get dropped every time the stat is repeated.

It measured qualification, not revenue. Getting someone on the phone and getting them to buy are different jobs with different drivers, and the study only spoke to the first one.

And the mechanism was a phone call to a stranger, in the United States, in the mid 2000s, when people still answered unknown numbers.

What changed underneath it

Mobile carriers and handset software now label unknown numbers as suspected spam. A large share of calls from a number the buyer has never seen are never answered, and many are never even shown. The channel the study measured has degraded, which means the same behaviour produces a different result.

The buyer changed too. Someone filling in your form in 2026 has usually read reviews, looked at two or three sites, and asked an AI assistant a question about your category before you knew they existed. They are further down the road than the 2007 lead was. Being first to speak matters less when the shortlist is already written.

Here is the split that actually holds. The value of speed scales with the urgency of the need, not with the size of the deal. A burst pipe, a locked-out tenant, a failed compressor in a cold room: minutes decide it, and the study's logic still applies at full strength. A $400K fitout with a procurement process and three stakeholders: your reply speed signals competence, but nobody is choosing a builder because you rang back in four minutes instead of forty.

Measure your own in three weeks

You need four timestamps per enquiry, and most CRMs already hold three of them.

1. 1.Form submit or call received.
2. 2.First genuine human contact attempt. An automated email does not count. A templated SMS saying we will be in touch does not count either.
3. 3.First two-way conversation, meaning they replied or picked up.
4. 4.Qualified or disqualified.

Bucket every enquiry by the gap between one and two: under 5 minutes, 5 to 30, 30 to 120, 2 to 24 hours, over 24 hours. Then calculate two rates per bucket. Contact rate is how many reached a two-way conversation. Qualification rate is how many of those became real opportunities.

Segment by source and by hour of arrival before you read anything into it.

The confound that ruins most of these analyses

Fast responses are not randomly distributed. They go to the enquiries that arrive at 10am on a Tuesday, when your best person is at their desk. Slow responses cluster at 7pm, on weekends, and over public holidays. High-intent forms often get routed to senior staff who respond faster anyway.

So when the report says five-minute responses convert better, some of that is response speed and some of it is that better leads were always going to get faster answers. You are measuring your own routing rules.

Two ways to break it. The clean one is a holdout: for a set period, randomly assign a slice of enquiries to a deliberate 30-minute delay and compare like with like. Most owners will not stomach that, which is fair. The workable one is to compare only within the same arrival hour and the same form type. If five-minute responses still beat 30-minute responses among Tuesday 10am enquiries from the same page, you have something real.

What to build instead of a five-minute phone sprint

Staffing a human to answer every enquiry inside five minutes, across every hour you take enquiries, is expensive. For most mid-sized businesses it means either a dedicated role or an offshore answering service that cannot qualify anything.

The cheaper structure that performs close to it:

An instant acknowledgement that carries a real person's name and a specific time commitment. Not a line saying we have received your enquiry. Something closer to Sarah has your enquiry and will call you before 4pm today. Specificity buys patience.

A booking link in that same message. Let the buyer take the appointment themselves. This is the piece most sites still do not have, and it removes the phone tag problem completely for anyone willing to self-serve.

A human call inside the window you promised. Missing your own stated window costs more than a slow first response ever did.

An SMS as the second touch rather than a second call. Answer rates on a text from an unknown number are far better than on a call from one, and the reply gives you the two-way conversation you were chasing.

What to do on Monday

Pull the last 300 enquiries. Calculate your median response time and your 90th percentile. The median is the number you will quote in a meeting. The 90th percentile is the number your worst-served customers actually experienced, and it is usually four or five times worse.

Fix the tail first. The gap between a five-minute median and a three-minute median is worth almost nothing. The gap between a 90th percentile of two days and one of two hours is worth a lot.

Stop quoting a number from 2007 and start quoting your own.