The Five Minute Cliff: what happens to an enquiry after the first 300 seconds

Reply to a new enquiry within five minutes and you are up to 21 times more likely to qualify it than if you wait half an hour. By minute ten the odds have already dropped four-fold. Firms that made contact within the hour were nearly seven times as likely to qualify the lead as those an hour slower, and more than sixty times as likely as those that waited a day. That decline is a cliff at the five minute mark followed by a long shallow tail, and most professional services firms are standing at the bottom of it: in the largest audit on record, the average business took 42 hours to answer a web lead, and 23% did not answer at all.

This article sets out what the evidence actually says, where the widely repeated numbers came from, and the three situations where responding faster will not help you.

Within five minutes. A firm that replies inside five minutes is up to 21 times more likely to qualify the lead than one that waits thirty minutes (MIT and InsideSales lead-response study, 2007), and the reply has to be substantive rather than a generic acknowledgement to earn any of that.

It is well supported but frequently misattributed. The 21 fold qualification figure comes from Oldroyd's 2007 MIT and InsideSales study, not the 2011 Harvard Business Review article. The widely cited McKinsey speed to lead study does not exist.

Before the data, a correction that matters. The figures you have seen quoted everywhere, that contacting a lead within five minutes produces a 21 fold increase in qualification odds and a 100 fold increase in contact odds, are almost always attributed to a 2011 Harvard Business Review article. They do not come from it. They come from Dr James Oldroyd's 2007 MIT and InsideSales study, which analysed over 15,000 leads and more than 100,000 call attempts across six firms.

The 2011 HBR piece is a different and separately useful study. It audited 2,241 US companies and found firms responding within one hour were seven times more likely to qualify a lead than those responding later, and sixty times more likely than those waiting over 24 hours. Average response time was 42 hours. Twenty three percent never responded.

There is also a widely circulated "McKinsey speed to lead study." No such primary research exists. It is a phantom citation that propagates because nobody checks.

We flag this because the credibility of the argument rests on the numbers being real. Several are. Several in circulation are not.

The decline is best stated the way the two surviving studies measured it: as odds of qualifying the lead, by delay. Conversion percentages by interval circulate widely in this sector; none we checked traces to a dataset, so this table carries odds only.

An earlier version of this article carried a synthesised decay table here, with consultation booking rates and deal conversion percentages by interval, credited to a spread of vendor benchmarks. When we went back and read each source in that caption, most did not say what the table said, two turned out to publish no lead-response research at all, and the flagship conversion column traced to a single vendor's unauditable marketing page. The table is gone. The odds above are the figures that survive the reading, and they are enough: the cliff is real, and it does not need invented precision to be alarming.

Two mechanisms drive the shape. The first is attention. Someone submitting an enquiry is, at that moment, focused on solving the problem. Within minutes they have moved to something else. The second is concurrency. People rarely contact one provider. They contact several, and the firm that answers first sets the reference point everyone else gets measured against. By the time a day has passed, the buyer who contacted several firms has usually stopped comparing and started instructing.

The gap between what the evidence recommends and what firms achieve is not marginal.

Legal illustrates why a single number misleads. In the Hennessey mystery-shop the median reply was 13 minutes, which looks excellent, while 26% of the same firms did not reply at all: a quarter of the market answers almost immediately and pulls the median down, and a long tail answers late or not at all. In a secret-shopper study of 500 US law firms, only 33% replied to an email enquiry (Clio Legal Trends Report, 2024). That figure measures email alone, so a firm silent by email may still reply on another channel.

The interaction to notice is between acquisition spend and non response: a firm pays real money to generate an enquiry and then, in a quarter or more of measured cases, no one answers it. That is an operations leak being paid for out of the marketing budget, and the industry pages set out where each sector leaks. An earlier version of this paragraph priced the leak with a cost-per-lead figure credited to a report that, on reading, does not contain it; the argument stands without the invented number.

The honest answer is that no reliable cross-sector share exists: the widely quoted figures trace to vendors quoting each other. The measured data point we trust is from UK estate agency, where 91% of agents say buyer and seller enquiries arrive outside nine to five, and 28% work beyond those hours every day of the week (Nested survey, 2021).

That is the structural problem, and it is timing rather than effort. The one place we found it actually measured is UK estate agency: 91% of agents report receiving buyer and seller enquiries outside nine to five, and 28% now work beyond those hours every day of the week (Nested survey, 2021). The dental, legal and cross-sector after-hours shares that circulate alongside that finding arrive without a dataset attached, so they do not appear here; the pattern they describe is real, on the evidence of the people living it.

No team answers every enquiry on the spot, every hour, every day. Evenings, weekends, and busy stretches are exactly when good leads arrive and go unanswered. A rota covers more of the week, never all of it, and the hours it cannot cover are the ones where intent is highest.

The famous statistics, that four in five callers hang up without leaving a message and three quarters go straight to a competitor, trace to no dataset: we followed the citation chains and they end at unsourced marketing copy. What is safe to say is that an out-of-hours caller was a buyer acting on intent, and a voicemail box answers nothing.

We went looking for the datasets behind the voicemail folklore, because this article repeats nothing it has not checked. The famous 80% hang-up figure dead-ends at an unsourced magazine line from 2014 with no study behind it at any link in the chain. The 86% variant is credited to a firm whose actual published figure differs from it by an order of magnitude. The claim that the 5pm to 8pm window accounts for 62% of after-hours calls cites a study that does not exist. What remains true without any of them: a call that reaches voicemail out of hours was a buyer with intent, and in an unstaffed practice it is simply gone.

A conversion figure for after-hours callers used to sit here, attributed to CallRail telemetry; CallRail has published no such figure, so it is gone. The argument that survives is behavioural: someone ringing a firm at 8pm is calling on their own time, further through their decision, and choosing to act now. They are also the enquiries most reliably lost.

The evidence has boundaries, and pretending otherwise is how vendors lose credibility.

Only when substantive. A generic acknowledgement is a receipt, and buyers read it as one. Replies that answer the question, qualify intent, or offer a bookable appointment are the ones that earn the speed advantage the research measures.

That is the first boundary. An instant "thank you for your submission, someone will contact you shortly" is a receipt, and buyers read it as one. Automated speed only earns anything when the reply does something: answers the actual question, qualifies intent, or offers a bookable slot. Research quantifying exactly how quickly a reader spots a template does not exist as far as we can find, so we make the point as judgement rather than dress it as measurement.

Switching channels squanders the gain. Mirror the channel the enquiry arrived on: answer a text by text and a web form with an emailed answer or a bookable slot, and treat an unsolicited phone call in reply to a form as friction rather than service. The reply-rate percentages often quoted for this point trace to a marketing article whose own source pages no longer exist, so the advice stands on its logic rather than on borrowed numbers.

High value advisory work behaves differently. In categories running under five enquiries a month, such as corporate transactions, restructuring, or ultra high net worth advisory, an instant automated reply signals commoditisation. Those buyers want senior attention and jurisdictional judgement, not velocity. Manual outreach remains the correct approach.

A fourth boundary worth stating: faster response cannot rescue an uncompetitive offer. Automating follow up on a proposition that does not convert simply increases the cost of not converting.

The evidence supports a narrow, specific conclusion rather than a general one. Speed matters enormously in the first five minutes, matters moderately for the first hour, and matters very little afterwards. It matters most for enquiries arriving when nobody is at the desk. And it only matters when the response carries substance rather than acknowledgement.

For a mid sized practice, the cost is worth estimating rather than quoting. In CallRail's 2022 survey of 508 US law-firm employees, published by a call-tracking vendor and built on self-reported answers rather than measured intake, firms admitted to an average five-hour response time and put their own annual losses at 46 clients and around $200,000 (CallRail, 2022 Marketing Outlook for Law Firms). Treat the specific figures as what they are, a sector's self-assessment, and the method as the transferable part. The equivalent figure for your practice is a straightforward calculation from four inputs you already hold: enquiry volume, current non response rate, conversion rate, and average client value. Most firms have not run it.

The ROI calculator runs exactly that sum with deliberately conservative assumptions and shows its working, so the figure you leave with is yours rather than a vendor's. And note what the evidence does not reward: working harder. No rota holds a five minute line through an evening, a weekend, and a busy Tuesday, which is why the response that meets the research is a structural one. That is the job an AI employee is built to hold: a reply within the minute, qualified against your criteria, with a bookable slot attached, at whatever hour the enquiry lands.