The lead decays while you're still reading the email.
This is not our claim. It is the most replicated finding in inbound sales, it is nearly twenty years old, and it is the reason we built a system around a clock instead of a conversation. Here it is, cited properly, with its weaknesses stated.
What the data says
In 2007, Dr James Oldroyd examined three years of behavioural data across six companies that generate and respond to web leads — more than fifteen thousand leads and over one hundred thousand call attempts. He measured what happened to a lead as a function of how long the company took to pick up the phone.
Odds of qualifying an inbound lead by response time. Five minutes is the baseline. At ten minutes the odds fall roughly fourfold; at thirty minutes, twenty-one-fold; after twenty-four hours, around sixtyfold.
Between five minutes and thirty, the odds of reaching a lead drop a hundredfold and the odds of qualifying one drop twenty-one-fold. The fall from five minutes to ten alone is roughly fourfold. The curve is not gentle and it does not wait for your sales team to finish lunch.
Almost nobody does it
This is the part that should interest you more than the decay curve. The research has been public since 2007 and the overwhelming majority of companies still lose the race. Not because they disagree with it — because a human being cannot reliably answer an enquiry in under five minutes at two in the morning, on a Sunday, while already on another call.
A machine can. That is the entire argument. Not that AI is impressive, but that this specific, narrow, unglamorous part of the problem is one a machine is strictly better at than a person, and it happens to be the part where the money leaks out.
Read it honestly
These are odds ratios from observational datasets, not a randomised controlled trial. Companies that respond in five minutes may differ from slow ones in ways the data does not capture — better staffing, better lead sources, better products. The effect is large and consistent enough that we build around it, but we are not going to pretend the causal claim is airtight, and you should be suspicious of anyone who does.
One more thing. This study is routinely credited to Harvard and quoted as "five versus ten minutes." Both are wrong. It is MIT/InsideSales, and the comparator is five versus thirty. If a vendor cites it the common way, they have not read it.
The 7× and 60× hour-scale figures come from the separate Harvard Business Review analysis, The Short Life of Online Sales Leads (2011), covering 2.24 million leads.
Contemporary response-rate benchmarks (23% / 42%) are from published 2026 B2B pipeline surveys.