Speed to lead statistics: which numbers hold up, and how sales teams time their own

The speed to lead statistics that hold up state their method and sample, like Clay’s 2026 test, in which two thirds of 6,346 companies never replied to a demo request. The 5 minute rule comes from a 2007 study of phone calls at 6 companies, and no study has been found behind “78% buy from the first to respond”. To time your team’s replies, use Obsession: its AI customers fill in the form, open the chat and call the sales line at the hours the team picks, and time every reply.

Speed to lead is the time from a buyer’s enquiry, such as a demo request, to the company’s first reply. Sales leaders, revenue operations teams and founders quote its statistics to set reply targets and to argue for more sales development reps, routing tools or AI agents.

The figures they quote come from 7 studies published between 2007 and 2026, each measuring something different. Two of the most repeated come from no study at all. A team’s own number comes from timing its own form, chat and phone the way a buyer meets them.

Speed to lead statistics, graded

Each statistic is traced to the page that first published it, with that source’s date, the sample behind the figure and one of 3 grades.

• Holds up: a public source states the figure, how it was measured and on how many companies, and the figure is quoted the way the source states it.
• Directional: the source is real, but its sample, method or age limits what the figure says about a sales team today.
• No traceable source: no study behind the figure has been found.

Where the 5 minute rule comes from

The 5 minute rule comes from the Lead Response Management Study. InsideSales.com, the sales software company now called InsideSales, ran it with James Oldroyd, then a faculty fellow at MIT’s Sloan School of Management. Oldroyd and InsideSales.com’s chief executive, Dave Elkington, presented it on 16 October 2007 at MarketingSherpa’s B2B Demand Generation Summit, and InsideSales.com announced the findings the same week.

The study’s own page describes 3 years of data from 6 companies that took leads through web forms. It covers over 15,000 leads and over 100,000 call attempts, all logged in InsideSales.com’s own system. It found that the odds of reaching a lead by phone were 100 times higher when the first call came within 5 minutes than at 30 minutes. The odds of qualifying the lead were 21 times higher.

Both figures describe phone calls. A contact meant a call that reached a live person for a set number of seconds, and each company used its own definition of a qualified lead. The study’s page states, “This study did not address close ratios.”

The rule is often credited to Harvard. The link is a later article by Oldroyd, Elkington and a Harvard Business School researcher, published by Harvard Business Review in 2011.

What the Harvard study found

The Short Life of Online Sales Leads appeared in the March 2011 issue of Harvard Business Review (HBR). Its authors were James Oldroyd of Brigham Young University, Kristina McElheran of Harvard Business School and David Elkington of InsideSales.com.

The authors audited 2,241 US companies, sending each a test lead through its website and timing the response. SmartCompany’s report of 1 July 2011, which quotes the article, gives the results.

• 37% replied within an hour, 16% within 1 to 24 hours and 24% after more than 24 hours.
• 23% never replied.
• Among companies that replied within 30 days, the average first reply came after 42 hours.

A separate analysis in the same article covered 1.25 million leads received by 42 US companies, 29 selling to consumers and 13 to businesses. Companies that tried to reach a lead within an hour were nearly 7 times as likely to qualify it as those that tried an hour later. Against companies that waited 24 hours or longer, they were over 60 times as likely. Qualifying meant reaching a key decision maker for a meaningful conversation.

HBR’s article opens in full to subscribers. InsideSales.com’s 2014 report gives the same 37%, 24% and 42 hour figures and cites the article, and Expertise.ai notes that the 42 hours is often misquoted as 47.

Figures with no traceable source

Two of the most repeated speed to lead statistics trace to no study anyone has found.

• 78% of customers buy from the company that responds first. Expertise.ai’s statistics page, updated in July 2026, traces it to a Lead Connect survey with no published method or original report.
• 35 to 50% of sales go to the vendor that responds first. It is usually credited to InsideSales.com. EmailAnalytics, which had cited a 2014 InsideSales report for it, says the share has no traceable primary source and that the report no longer publishes a method it can check.

Faster contact does raise the odds of a qualified conversation, as the 2007 and 2011 studies found. These 2 figures can’t be checked, so neither belongs in a business case or a board deck.

What the 2024 and 2026 tests show

RevenueHero requested a demo from 1,000 B2B SaaS companies, most with 51 to 500 staff, in a study published in March 2024. Of the 1,000, 365 replied and 635 never did. The average reply took 1 day, 5 hours and 17 minutes. The fastest 172 companies replied instantly, in about 2 minutes on average, and 31 more replied within the hour.

Of the companies that replied, 60% sent an automated first response. Automated first responses took 17 hours and 20 minutes on average, and replies written by a person took 2 days, 3 hours and 11 minutes. RevenueHero adds that 31% of the companies that never replied used an enrichment tool, software that looks up the company behind a form, so some may have screened its request out.

Clay published its test on 22 September 2026. Agents built in Clay filled in 6,346 demo and contact forms at companies in 100 countries and sent no follow up. Every form came from the same made-up buyer, the founder of an AI startup.

Email replies came to 2,013 forms, 32% of the total. Only 100 companies phoned, about 5% of those that replied, and their calls arrived 34.5 hours after the form on average. About 7% of the forms got a reply from a salesperson, and 5% of replies arrived within 5 minutes.

Clay puts its average speed to lead at 15 to 24 hours depending on the cut, or 26 to 40 hours for replies written by a person. It lists its own caveats: a missing reply may have been a deliberate choice, spam filters may have blocked some forms, and an AI judged which replies were automated. Bigger companies replied faster on average and automated more of their first replies.

Optifai, a sales software company, reports an average of 47 hours on a page updated in April 2026, with 23% of companies responding within 5 minutes. Its figures come from CRM and marketing automation data at 939 B2B companies, from the second quarter of 2025 to the first of 2026. It times from the lead’s creation in the CRM to the first sales activity, and its page doesn’t say how the companies were chosen.

Every audit since 2011 found a large share of companies that never answered the test lead. The share was 23% in HBR’s audit, 47% in InsideSales.com’s 2013 audit, 63.5% in RevenueHero’s and about two thirds in Clay’s. Each used its own method, so the rise between them isn’t a like-for-like trend.

Why the numbers disagree

Published averages run from hours to days because each study starts its clock in a different place, counts different replies and tests at different hours.

• Where the clock starts. The audits start it when the buyer sends the form. Optifai starts it when the lead is created in the CRM. HubSpot’s lead response time report starts it when a contact is assigned to a user, and starts it again when the contact is reassigned.
• What counts as a reply. HubSpot counts the owner’s first email, call, chat, task marked in progress or done, or meeting. RevenueHero timed automated first responses and a person’s replies separately, and Clay had an AI sort the two.
• When the test runs. InsideSales.com’s 2013 audit sent its test leads between 08:00 and 17:00 local time, so it measured no evening or weekend leads. RevenueHero and Clay don’t say when they sent theirs.
• Who asks. Clay sent the same made-up founder to every company, so a company that rightly judged him a poor fit counts as a non-reply.

A team’s own report can show 32 minutes while the buyer waited more than 4 hours, because its clock starts later and a lead that never became a record never counts. A test from the buyer’s side starts the clock when the buyer presses send.

How sales teams time their own

A labelled test lead goes through the team’s own form, chat and phone at several times of day, after hours included, and each reply is timed against a target set first.

1. List every way in. Each demo and contact form, the site chat, the sales line and the published sales email address, with the owner of each step behind it. A pass: the list holds every way a buyer can reach sales.
2. Set the target first. For example, an owner in the CRM within 5 minutes and a first reply from a person within 5 minutes in staffed hours. Outside them, a reply by 09:30 the next working day. A pass: sales, marketing and revenue operations agree the numbers before the first test.
3. Label every test lead. A hidden form field, a tagged email address or a fixed company name lets the CRM keep test leads out of attribution and quota. A pass: every test lead appears in the CRM marked as a test.
4. Send it at several times. Spread 5 tests a week across the day and the week, for example at 09:00, 13:00 and 17:00 on different weekdays, at 21:00 on a Friday and once on a Sunday. Each test asks a question a real buyer would ask. A pass: every time slot meets its target.
5. Use each way in as a buyer does. Fill in the form, open the chat and call the sales line, leaving a voicemail if nobody answers. A pass: the reply answers the question that was asked.
6. Set the window, then compare. Read the results at 4, 12, 24 or 48 hours, or keep watching as long as the team wants. Call a reply missing only once that window closes. Then set the log beside what the CRM recorded. A pass: every test lead has an owner in the CRM and a reply inside the window.
7. Test again after every change to routing, the CRM, a form or the site. A pass: every path that passed before still passes.

A team with 5 ways in, each tested 5 times a week, runs 25 tests a week before any retest. Each one needs an inbox and a phone watched for days, and reps soon learn the tester’s name.

How AI customers time a team’s replies

An AI customer is an AI agent that becomes a real customer of a company, with its own inbox, phone number and browser, and keeps what the company sends and shows it as evidence. For a sales team, the evidence is the reply itself, with the time and channel it came on. How AI customers work shows each step.

AI customers act as the test leads. Each has its own inbox and phone number. It fills in the form, opens the chat and calls the sales line at the times the team picks, evenings and weekends included. It keeps every reply with its time and channel. Whether it also says it’s a test is the team’s setting.

Obsession makes AI customers that sign up to any company and keep every email, offer and price it sends, so you can find your revenue leaks, watch your competitors and prospect with proof. Lead leaks, its speed to lead test, times the first reply to every enquiry on the form, chat and phone the team owns. It shows each lead nobody answered and drafts the routing fix in the CRM the team connects, for the team to approve. The team behind Obsession has mystery shopped 1,500 ecommerce stores as a new customer since September 2026.

The same AI customers check the team’s own AI agents. Sales agent check asks the AI agent in the site chat about prices, stock and delivery every day and checks each answer against the price list. Outbound agent check joins the lists of an AI SDR, an AI agent that writes to prospects for the team, and checks every email and call it sends against the team’s rules. AI customer service agents vs AI customers sets out who each agent works for.

Each AI customer says it’s an AI when it talks to a person and never gets round a CAPTCHA, a puzzle that blocks automated visitors, or a bot block. Where a form puts a CAPTCHA in front of a step, the report names the step it couldn’t test. What every AI customer will and won’t do sets out the full list. Why a lead went to the wrong owner sits in the router and the CRM, where the team’s own tools show it.

Ways to time your replies, compared

Obsession makes AI customers, so every other option is described from its provider’s own pages as of 1 November 2026. Tools that act as a customer of a business in other ways, such as calling staff or testing a chatbot, are compared in AI mystery shopping tools. Tools that watch competitors, such as email archives and page monitors, are set beside Obsession in Obsession and other tools.

If your team needs to know the first reply a buyer gets on every form, chat and sales line, at the times it picks, evenings and weekends included, use Obsession. A CRM report suits coaching reps once a lead has an owner. A form monitor catches a form that breaks outright, and a test by hand suits a spot check after a change.

Questions about speed to lead statistics