Nearly every speed to lead statistic in circulation traces to one source: the 2007 Lead Response Management study led by Dr. James Oldroyd, then at MIT. Calling a web lead within 5 minutes instead of 30 makes contact 100 times more likely and qualification 21 times more likely. Response speed is the strongest conversion lever most teams never measure.
What did the 2007 Lead Response Management study actually measure?
The study is narrower and more rigorous than the pages citing it suggest. Oldroyd analyzed when companies called their web-generated leads and what happened next. He measured one variable with precision: the elapsed time between a lead's creation and the first call attempt. He mapped it against two outcomes: whether the call reached the lead, and whether the conversation qualified them. Not close rates. Not revenue. Contact odds and qualification odds. That restraint is what made the finding durable.
The dataset behind the finding: three years of response data across six companies, covering more than 15,000 leads and over 100,000 call attempts.
The odds of making contact with a lead called within 5 minutes versus 30 minutes are 100 times higher. The odds of qualifying that lead are 21 times higher.
Read those numbers as a decay curve, not a benchmark. A lead's attention starts evaporating the moment the form is submitted. At five minutes they are still at their desk, on your site, in the mindset that made them reach out. At thirty minutes they are in a meeting, on a competitor's site, or gone.
Why does a study from 2007 still dominate this query?
Three reasons. First, it measured odds ratios rather than era-specific benchmarks. Channels change. The decay of human attention does not. Second, nothing comparable has replaced it in public. The study was conducted with InsideSales.com, a sales software company with an obvious stake in the outcome, but it published its dataset and method, and most successor studies are vendor surveys that publish neither. The work with real numbers keeps winning citations by default. Third, almost nobody citing it has read it. The 100x figure circulates stripped of what it measures, which is contact odds, and gets misquoted as a revenue multiplier. Precision about the claim is now rarer than the claim itself.
What has changed since 2007?
The study measured phone calls to web form leads. Today the same buyer arrives through chat, text, ad clicks, and direct messages, at midnight, on weekends, in parallel. Buyer expectations hardened alongside: consumer apps trained everyone to expect an answer in seconds, so a next-morning callback now reads as indifference. The biggest change is the constraint itself. In 2007, responding inside five minutes meant staffing a call floor around the clock. Today an autonomous system can engage every inbound in seconds at any hour. The capability exists. Adoption has not caught up.
A Harvard Business Review audit of 2,241 U.S. companies found the average first response to a web lead took 42 hours. Only 37 percent responded within an hour. 23 percent never responded at all.
Two decades of tooling later, the window is still being lost. The reason is structural, not motivational. Response time is a property of the system, not of the salesperson, and most businesses have never engineered the system.
How do you apply the odds to a real pipeline?
The study's numbers are odds ratios, so apply them to your own funnel instead of borrowing a benchmark. The exercise takes one afternoon.
- Export the last 90 days of inbound leads with two timestamps: creation and first human or system touch.
- Compute the response-time distribution, not the average. One weekend backlog hides inside an average and dominates a distribution.
- Bucket leads by response window: under 5 minutes, 5 to 30 minutes, 30 minutes to 24 hours, beyond 24 hours, and never touched.
- Compare contact and qualification rates across the buckets. Your own data should reproduce the study's decay shape.
- Multiply the leads falling outside the 5-minute window by your qualification gap and average deal value. That figure is what slow response costs you per quarter.
Then run the arithmetic on a concrete shape. Suppose 200 leads arrive in a month and 150 of them get a first touch after the 30-minute mark. The study's odds say contact on those 150 is two orders of magnitude less likely than it would have been inside 5 minutes, and you paid identical acquisition cost for every one. For a human team the 5-minute window is not a stretch goal, it is a structural impossibility. Nights, weekends, lunch, and two leads arriving at once all break it. This is an engineering problem, not a discipline problem.
Why is speed without governance reckless?
The obvious response to these numbers is to automate everything and answer instantly everywhere. That is half right. Speed without qualification logic fills your calendar with meetings a filter should have caught. Speed without approval gates means a system can message someone who never opted in, publish a wrong answer, or commit spend, all at the same machine speed it was praised for. Automation amplifies whatever it executes, including the mistake. The fix is not to slow down. It is to separate classes of action. Capture, scoring, routing, and follow-up with warm contacts are bounded and reversible, so they should run at full speed with nobody babysitting them. Net-new outreach, anything that spends money, and anything public are expensive to get wrong, so they should pause for a human who sees the full context. The pause costs minutes. The unforced error costs the trust the entire system runs on.
Speed to lead is physics, not a tactic. The window is five minutes, so our systems engage in seconds and never sleep. But the same system stops before it spends money or messages anyone brand new, shows you the full context with a recommended decision, and waits for you. Fast where fast is safe. Deliberate where a mistake is expensive. That pause is why our systems are still running when others got switched off.
What should you demand from a speed to lead vendor?
We build a Lead & Growth Engine that engages every inbound in seconds, scores it against what your best customers look like, and pauses for human approval before any net-new outreach or spend. Hold any vendor you evaluate to that same standard. Ask for a measured response-time distribution against your own baseline, and a written list of the actions the system will never take without you. If either answer is vague, so is the system.