Search for “lead response time statistics” and the first page is a wall of numbers. Forty-seven data points. Thirty-two statistics. Every one of them confident, every one of them formatted as a bullet, and almost none of them carrying a link to anything you can read.
We are about to publish a lot of writing in this category, and we sell a product whose entire promise is that it will not state a figure it cannot back up. Repeating a statistic we had not checked would have been the fastest possible way to be a hypocrite. So before writing anything else, we spent a day trying to find the original source of the ten claims our industry repeats most.
Four survived with their primary source intact. Two should be retired outright. The rest need caveats that nobody printing them has been printing.
The results
| The claim | Verdict |
|---|---|
| “Respond within an hour and you are 7× more likely to qualify the lead”, HBR, 2011 | Real, and misquoted almost everywhere |
| “The MIT study”, 21× and 100×, 2007 | Real study, not an MIT study, phone calls only |
| “78% of customers buy from the company that responds first” | No source exists |
| “50% of sales go to the vendor that responds first” | No source exists |
| “Only 27% of leads are ever contacted” | Vendor’s own unpublished research |
| “Responding in one minute lifts conversion 391%” | Real, 2012, and it measured phone calls |
| Drift’s response-time benchmark reports | Original URLs are dead |
| Baymard: 70.22% cart abandonment | Clean. Transparently sourced |
| Gartner: 87% want a human option | Clean. Sample and dates published |
| An independent mystery shop since 2020 | Exactly one exists |
The five-minute rule is not in the paper everyone cites
The most-repeated idea in this industry is that lead quality falls off a cliff after five minutes, and the citation attached to it is almost always The Short Life of Online Sales Leads, Harvard Business Review, March 2011.
That paper is real. It is by James Oldroyd, Kristina McElheran and David Elkington, it audited 2,241 US companies and analysed 1.25 million leads from 42 of them, and its findings are strong. Firms that made contact within an hour were nearly seven times as likely to qualify the lead as those that tried an hour later, and more than sixty times as likely as those who waited a day. It also found that 23% never responded at all and that the average response time was 42 hours.
The number five does not appear anywhere in it. The paper measured hours.
The five-minute threshold comes from a different piece of work, and the citation has been swapped so many times that the two are now permanently confused. If you have ever quoted HBR for the five-minute rule, and we nearly did; you have cited the wrong paper.
The “MIT study” is not an MIT study
The five-minute figure comes from the Lead Response Management Study, presented in October 2007. It is universally called “the MIT study”, and that framing does a lot of work it has not earned.
There is no original paper at any MIT or academic address. What circulates is a seven-page executive summary hosted on vendor domains and file-sharing sites. It was never peer reviewed, and it was never published by MIT as an institution, the label rests entirely on one author’s affiliation at the time. The underlying data came from InsideSales.com’s own system: roughly 15,000 leads across six companies, and more than 100,000 call attempts.
Two things get lost in the retelling, and the second one matters enormously if you are reading this because you are thinking about live chat.
The direction is reversed. The original says the odds drop, “the odds of qualifying a lead if called in 5 minutes versus 30 minutes drop 21 times”. Those are odds ratios between two specific points in time. They are routinely restated as an unqualified “21× more likely to qualify”, which is a different and much larger claim.
It measured phone calls. Leads arrived by web form and were then called. The study says nothing whatsoever about how fast you should answer an email or a chat message. It is nonetheless the most-cited statistic in chat software marketing, including, until we checked, very nearly ours.
Two numbers that do not have a source at all
Two claims turn up on almost every page in this category:
- “78% of customers buy from the company that responds first.”
- “50% of sales go to the vendor that responds first.”
The 78% figure is attributed, everywhere, to “a Lead Connect survey”. There is no published report, no methodology, no sample size and no date. The pages citing it link to each other, or to nothing, here is one of them. We could not find a single page that reproduces the survey itself.
The 50% version traces back to a 2015 blog post that hedges its own headline, “a reported 35% to 50%”, while pointing at a CEB and Google white paper that is about where buyers are in their journey, not about how fast anyone replied.
We are not saying these numbers are wrong. We are saying nobody can show you where they came from, which means nobody can tell you what was measured, or when, or of whom. They will not appear anywhere on this site.
The 391% is real, and narrower than you think
This one surprised us. “Responding within one minute increases conversions by 391%” is usually printed without a source, and it turns out there is one: Velocify’s The Ultimate Contact Strategy, published in 2012, not the 2013 or 2014 usually given. It drew on around 3.5 million leads across 400-plus companies, which is a serious dataset.
Three caveats travel with it and none of them usually do. Those were Velocify’s own customers, weighted toward phone-heavy industries like mortgage and insurance. It measured placing a phone call within one minute, not “responding”. And the report’s own text says “nearly 400%”, the precise-sounding 391% is read off a chart. Velocify no longer exists as a company, and the report survives only because other people mirrored it.
The company that made response time famous has deleted its own numbers
Drift built a category on this idea. Its secret-shopper study, 433 B2B SaaS companies, published February 2017, not 2018 as usually cited , found that only 7% responded within five minutes and 55% did not respond within five business days.
Every original URL for it now redirects to a product page following the acquisition. The study survives on a blog and a Medium post. Worth noting too, since almost nobody does: of the ten fastest responders, the report observes that all of them used live chat. That finding is in a report published by a live chat company, and it is fine to say so.
What is actually independent
One thing genuinely qualifies. Pied Piper has run third-party mystery shops in the automotive industry for over a decade, submitting real enquiries about real vehicles and scoring what comes back across email, phone and text over 24 hours. The 2024 study submitted 1,535 enquiries across 18 dealer groups; a 2025 study of compact tractor dealers shopped 726 dealerships and found customers were ghosted 20% of the time.
Two more are clean but should be labelled honestly. Conversica shopped 100 companies over 22 days in 2023 and found one in four never responded to an inbound lead at all, good methodology, run by a company selling AI lead follow-up. Baymard’s 70.22% cart abandonment figure aggregates 50 studies and says so, with a last-updated date on the page, which is more transparency than anything else in this audit. And Gartner, in August 2026, found that 87% of customers say a company using generative AI for service must provide a route to a human, 3,566 respondents, fielded February to March 2026, sample size published.
How to check one yourself, in about four minutes
- Find the primary source. Not the page quoting it. Follow the citations back until you hit a document or a dead end. A dead end is an answer.
- Check what was measured. Phone calls are not emails. Emails are not chat. A great deal of chat marketing rests on a study about outbound calling.
- Check the sample. Whose customers were they? A vendor’s own client base is not the market.
- Check the date. Several of these are from 2007 and 2012, before most of the behaviour they are used to explain existed.
Why we did this before writing anything else
We are a new company. We have no case studies, no logos and no testimonials, and we have said so on our own homepage rather than inventing them. That leaves one honest way to earn a reader’s trust, which is to be the people who check.
It also happens to be the same discipline the product runs on. Ampa cannot state a price that is not in its client’s approved data, because every message it sends is checked against that data before it goes out, a matcher in code, not a line in a prompt. Writing a blog post to a lower standard than the software would have been an odd thing to do.
Questions people ask about this
Is the five-minute rule real?
Partly. The five-minute threshold comes from the Lead Response Management Study presented in 2007, which measured roughly 15,000 leads across six companies and found the odds of qualifying a lead drop about 21 times between calling at five minutes and calling at thirty. It measured outbound phone calls, not email or web chat. The Harvard Business Review paper it is usually credited to measured hours, not minutes.
Was the MIT lead response study actually done by MIT?
No. There is no original paper at any MIT or academic address, it was never peer reviewed, and MIT never published it as an institution. What circulates is a seven-page executive summary hosted on vendor websites. The label comes from one author's university affiliation at the time. The underlying data came from InsideSales.com's own system.
Where does '78% of customers buy from the company that responds first' come from?
Nowhere that can be checked. The figure is attributed everywhere to a Lead Connect survey, but no published report, methodology, sample size or date exists for it, and the pages quoting it cite each other rather than a source. The same is true of the 50% variant.
What did the Harvard Business Review 2011 study actually find?
It audited 2,241 US companies and analysed 1.25 million leads from 42 of them. Firms that made contact within an hour were nearly seven times as likely to qualify the lead as those trying an hour later, and more than sixty times as likely as those waiting a day. It also found 23% never responded at all, and an average response time of 42 hours.
Which lead response statistics can I safely cite?
Four hold up with a primary source: the Harvard Business Review 2011 study, if you quote its hour-based findings rather than a five-minute rule; Velocify's 2012 figure, if you note it measured phone calls placed to the vendor's own client base; Baymard's cart abandonment average, which states that it aggregates 50 studies; and Gartner's 2026 finding on human escalation, which publishes its sample size and field dates.
How do I check a statistic myself?
Follow the citation back until you reach a document or a dead end, because a dead end is an answer. Then check what was actually measured, since phone calls are not emails and emails are not chat. Check whose customers were in the sample, because a vendor's own client base is not the market. And check the date.
The short version
Before you cite a statistic in this category, follow it back to the document. Four of the ten most-quoted numbers survive that walk intact, two do not exist, and the most famous one measured something other than what it is used to prove.
We built the same rule into the software.
Ampa cannot state a price that is not in your approved data, because every message is checked against that data before it is sent. Ask the agent on this page what Ampa costs; it will refuse to put a figure in the chat window, and email you a real one.
Start free trial → or ask the agent something hard first