Contents
Speed to lead is the time between a prospect asking to hear from you and your first reply. The largest recent benchmark suggests most B2B companies are slow, and many never answer at all. When Clay's agents submitted 6,346 demo and contact forms, two-thirds of companies never replied, 96% never phoned, and only about 7% of forms got a reply from a salesperson. Clay published the results on 22 September 2026.
This piece covers what the test found, where its numbers may be off, and how a GTM engineering team can fix speed to lead.
How was the test run?
Alex Lindahl, Clay's Creator-in-Residence, built the test in a single Clay table. It started with more than 10,000 target domains. An AI agent found each company's demo or contact form, a browser agent filled it in using one consistent buyer identity, and the system logged the time of any email or call that came back. No follow-up was sent.
The 6,346 forms that were submitted came from companies in 100 countries, across software, professional services, manufacturing, retail and other sectors.
What are the speed to lead benchmarks?
The headline numbers from Clay's report:
1/ Only 2,013 forms, 32%, got any email reply.
2/ Just 100 companies phoned, and those calls arrived on average 34.5 hours after the form was sent.
3/ Of the 2,013 replies, 64% were automated. 732 came from a person, and 445 of those were a salesperson trying to start a conversation. That is 7% of all forms.
4/ Average speed to lead was 15 to 24 hours depending on how the data was cut, and 26 to 40 hours for human replies.
5/ 5% of replies arrived within five minutes.
6/ 75.8% of human replies were fully personalised. Of 1,281 automated replies, four were.
7/ Bigger companies were faster on average and automated more of their first replies.
Clay's summary of the pattern: "Hardly any managed to be both quick and personal." Companies either sent a template within seconds or a personal reply a day later.
What does a slow reply cost?
Take an illustrative company that receives 120 demo requests a month. At the benchmark reply rate of 32%, about 38 get any answer and 82 hear nothing. At the 7% salesperson rate, about eight start a real sales conversation.
Now suppose a third of all requests are genuine buyers and the company wins a quarter of the buyers it actually speaks to. Replying to all 120 with a person means talking to 40 buyers and winning about 10. Replying like the benchmark means talking to fewer than three of those buyers and winning less than one. The leads were paid for either way.
Where might the numbers be off?
Clay is open about the limits, and they matter.
1/ Every form used the same buyer: the founder of an AI startup. Some companies may have decided, correctly, that he was not a fit, or suspected a bot. The data cannot separate that from plain silence.
2/ Spam and bot filters may have caught some of the agent's email addresses, which would understate replies.
3/ The split between human and automated replies was judged by AI, which can misread a well-written template.
Even allowing for all three, the gap between how fast buyers expect an answer and how fast they get one is wide.
What did the best responders do differently?
One detail stands out. More than 50 companies ignored the phone number on the form and called the buyer's real number, and around 20 called his colleagues. Those teams had enriched the lead before dialling. They knew who had asked before they picked up the phone.
That is the GTM engineering fix in miniature: research the lead automatically, before a rep sees it.
How do you fix speed to lead?
Clay's recommendations, with the reasoning behind each:
1/ Reply to everyone. "Coverage is a bigger problem than speed," Clay wrote. An automatic acknowledgement on every submission is the minimum.
2/ Give disqualified leads a next step. A recorded demo, a self-serve trial, a pricing page. Silence is the worst answer, and today's poor fit may be a buyer at their next company.
3/ Enrich and route before the rep opens the lead. When company size, role and fit are already attached, the first human reply can go out quickly and with context.
4/ Call your best-fit leads the same day. With 96% of companies never phoning, a same-day call is rare enough to stand out.
How fast should you reply?
There is no single right number, but the benchmark sets a low bar. If your human replies land in under an hour and your best-fit leads get a call the same day, you are ahead of almost everyone in this sample. The simplest test is the one Clay suggests: fill in your own demo form from a personal address and time what happens.
A note on Peach Data
Peach Data builds go-to-market engineering systems for sales and marketing teams: the data foundation, the workflows and the agents behind outbound and inbound. We started as a verified data business, so every build starts with the data layer. peachdata.co.uk
