Speed to Lead: Why Fast Responses Only Pay Off on Genuine Leads

Speed to lead is the most studied number in sales, and possibly the least acted on. The evidence goes back nearly two decades and it all points one way: get to a web enquiry within five minutes and the odds of a conversation are on your side; wait an hour and they fall off a cliff. What the conference decks rarely mention is the denominator. When a share of your form fills are bots, competitors and outright fakes, your response time is being measured against enquiries that were never going to buy. Speed applied to junk is wasted motion, and the wait it creates for genuine buyers is the real cost.

What is speed to lead, and how fast is fast enough?

Speed to lead is the elapsed time between a prospect submitting an enquiry and your first meaningful attempt to contact them. The research consensus is blunt: minutes, not hours, with the odds of contact falling more than tenfold inside the first hour. A 2011 Harvard Business Review audit put the average company at 42 hours.

The decay curve is old news to anyone who has run an inbound team, but the numbers hold up better than most sales research. In 2007, Dr James Oldroyd, working with call data supplied by InsideSales.com, examined three years of records across six companies: more than fifteen thousand leads and over one hundred thousand call attempts. The finding everyone quotes is the hundred-fold drop in contact odds between a five-minute callback and a thirty-minute one. The finding that matters operationally is that qualification odds fall more than sixfold inside the first hour.

Harvard Business Review took a different route in 2011. Rather than mine call logs, James Oldroyd, Kristina McElheran and David Elkington submitted test enquiries to 2,241 US companies and waited to be contacted. Twenty-three per cent never replied at all. Those that did respond took 42 hours on average, and those that tried within the first hour were markedly more likely to reach a decision maker. The authors' summary still reads as current: most companies are not responding nearly fast enough.

Stat block: what the response-time research actually says

"The odds of calling to contact a lead decrease by over 10 times in the 1st hour. The odds of calling to qualify a lead decrease by over 6 times in the 1st hour." — 2007 Lead Response Management study (Dr James Oldroyd, with InsideSales.com call data)

"The average response time, among companies that responded within 30 days, was 42 hours." — Harvard Business Review, "The Short Life of Online Sales Leads", 2011

"Firms that tried to contact potential customers within an hour of receiving a query were nearly seven times as likely to qualify the lead … and more than 60 times as likely as companies that waited 24 hours or longer." — Same HBR study, 2011

"When we requested a demo on 1000 B2B websites, we received only 365 responses in total with an average response time of 1 day, 5 hours, and 17 minutes." — RevenueHero test of 1,000 B2B SaaS companies, 2024

Has any of this improved? Not obviously. RevenueHero's 2024 test is the current picture, and its own caveat matters: many non-responders used enrichment tools and may have silently discarded the test lead as a poor fit. Even allowing for that, a world where most genuine hand-raisers wait a day is one where five-minute discipline still wins deals.

Why does a fast response only pay off on genuine leads?

Because response speed converts intent, and a fake lead has none. Every minute a rep spends on a bot submission is a minute a genuine buyer waits, and the decay curve runs on the genuine buyer's clock, not the fake's. The response window is zero-sum: junk quietly taxes the enquiries you paid to win.

The first failure mode is arithmetic. Reps work queues in rough order of arrival, so a Monday-morning burst of forty bot submissions pushes every genuine enquiry behind it. Your SLA clock keeps running on the real leads whilst the team rings numbers that were never going to answer. A five-minute promise that survives a junk flood was measured on a quiet day.

The second failure mode sits in your automation. Round-robin assignment spreads fakes evenly across the team, so everyone's numbers degrade together. Lead scoring learns from whatever you feed it, and junk conversions are cheap conversions. Lunio's analysis of invalid traffic makes the point plainly: fake leads are usually much cheaper than real ones, which "can create a negative feedback loop within your automated campaign, whereby the algorithm continues to seek out more junk conversions to make your acquisition costs lower". The short version of how junk leads train ad platform optimisation: the ad account learns from the leads you accept, not the leads you wish you had.

The third failure mode is human. Reps who spend their mornings dialling dead numbers stop treating the alert as urgent, because nine times out of ten it isn't. Cynicism sets in within weeks, and it looks identical to laziness in the metrics: first-response times drift out and follow-ups get skipped. No dashboard labels this as junk-induced, which is why teams blame culture when they should blame the queue.

Nor is this a fringe problem. Lunio's 2024 Wasted Ad Spend Report found "more than two-thirds (69.1%) of performance marketers report fake leads from paid media campaigns". That is the share of marketers who have seen the problem, not the share of leads that are fake. Even so, the safe assumption is that some of your queue is fake too, concentrated wherever you spend most.

How do you keep fake leads out of the fast lane?

Validate at the form, in real time, before anything reaches a human queue. Email, phone, behavioural and IP checks catch most junk at submission, so only genuine enquiries enter the fast lane. Then clean the backlog once, and only then measure response time, because a benchmark set against junk is a benchmark you cannot trust.

Real-time validation is the piece most teams skip, usually for fear of slowing the form down. LeadShield.ai blocks fake, fraudulent and AI-generated leads at the web form itself, from one JavaScript snippet that takes about sixty seconds to install. Detection runs across email validation with disposable-domain, MX and typo checks, phone validation, behavioural bot detection including honeypots and fill-time analysis, and IP reputation. Every blocked lead is logged with a plain-English reason and exportable as CSV.

On Professional plans and above, a valid verdict reaches your own stack, a Slack or CRM alert for instance, the moment a genuine lead passes, with HMAC-signed webhooks doing the delivery. LeadShield becomes the rail your speed-to-lead promise runs on: genuine enquiries arrive pre-checked, fakes never enter the queue. Plans start at $39 a month and the 14-day trial needs no card. To see the verdict flow on your own form, book a LeadShield demo.

The backlog deserves its own pass. Before you can measure response time honestly, you need to know how much of the existing CRM is worth calling. Every LeadShield plan, trials included, includes bulk CSV cleaning: upload a CRM export and get a verdict, risk score and plain-English reason per row, with 250 rows per upload on Starter, 1,000 on Professional and 5,000 on Enterprise. It runs deterministic checks only, with no behavioural analysis on CSV rows, so treat it as a sieve rather than a judgement on intent. If you would rather start by hand, there is a practical method for auditing a batch of 250 leads with no install.

Should you measure response time before or after cleaning the backlog?

After. A response-time benchmark taken before the clean-up is dominated by rows nobody should have called, so it will flatter you, mislead you, or both. Clean the backlog, agree what counts as a genuine enquiry, then measure the share of leads answered inside five minutes. Track the median and the slowest decile, not the average.

Keep the measurement simple. The share of genuine leads answered inside five minutes, published weekly, does more for behaviour than any dashboard of averages, because an average survives a day-long outlier and a five-minute promise does not. Pair it with a named owner for the first five minutes and a written rule for leads that arrive at 7pm. None of this is sophisticated. It is just easier to hold a discipline when the queue behind it is real.

One last honesty note, because this field attracts folklore. You will see the claim that the first vendor to respond wins some fixed share of deals, usually 78 per cent. No primary study behind that number can be traced; it circulates through round-up posts that cite each other. The verifiable findings are the decay figures above, and they are strong enough without decoration.

Frequently asked questions

How quickly should we respond to a new lead?

Within five minutes where you can, and inside the first hour as the hard limit. The 2007 Lead Response Management study found contact odds fall more than tenfold in that first hour, and HBR's 2011 study found a one-hour attempt was nearly seven times more likely to qualify the lead than waiting even an hour longer. If five minutes is not realistic for your coverage hours, fix the coverage before you fix the speed.

Is it true that 78% of buyers choose the vendor that responds first?

That figure has no traceable primary source, despite how often it is repeated. It is attributed across vendor blogs to a survey nobody can produce. The defensible numbers are the decay figures from the 2007 and 2011 studies, which make the same argument with evidence behind them.

Do fake leads really affect response time, or just conversion?

Both, and response time is the more insidious of the two. Junk consumes the response window by sitting in the queue ahead of genuine buyers, and it corrodes the routing and scoring rules that decide who gets called first. Conversion loss shows up in the funnel; response-time loss hides in a number measured against the wrong enquiries.

What is the cheapest first step?

Clean a sample of the existing backlog before buying anything. Upload a CRM export to a bulk cleaning tool, or work through a sample manually, and establish what share of recent enquiries were genuine. That one number tells you whether your response-time problem is a discipline problem, a junk problem, or both.