Buyer Wait Time, the metric specification

Full figure set: 62 B2B buyer response statistics.

Buyer Wait Time is the time between a buyer explicitly asking to reach a person and a person answering.

Everything else here exists to make that sentence unambiguous, so that two people measuring the same company independently arrive at the same number. This is a forward specification for anyone measuring their own company. Where it differs from how our own 2023 study was collected, the difference is stated.

The four rules

Rule 1: the clock starts on an explicit request for a human, not a visit

The start event is a buyer asking to speak to a person. It is not attention, interest, or presence on a website.

Browsing does not start the clock. Opening a chat widget does not start it, because opening a widget is an attempt to find out what is behind it rather than a request. A question that a bot answers to the buyer's satisfaction does not start it either, because the buyer got what they came for and never asked for a person.

What does start it: submitting a demo request or contact sales form, asking in chat to speak to someone, requesting a call back, or any equivalent action where the buyer has stated that they want a human.

Rule 2: the clock stops when a human answers

The stop event is a person responding to that specific request. An autoresponder does not stop the clock, nor does an acknowledgment of receipt, nor a bot. A calendar link sent automatically does not stop it either, because the buyer asked for a person and received a scheduling instruction.

The rules for deciding whether a given response was written by a person are on the email classification rulebook page. The governing test is whether the exact response could have been sent, unchanged, to anybody else who filled in the same form.

Rule 3: channel agnostic

Buyer Wait Time does not privilege a channel. Chat, email, telephone, text message or any other route counts. The clock stops on the first human response through any channel, including a channel different from the one the buyer used. A buyer who submits a web form and receives a telephone call from a person forty minutes later has a Buyer Wait Time of forty minutes.

Channel-specific metrics let a company report one route while the buyer waits on another.

Rule 4: closed and unanswered are different failures

A company that offers no path to a human is closed. That can be a defensible design decision. A self-serve product with no sales motion is not failing when it declines to staff a queue it never opened.

A company that offers the path, watches a buyer take it, and never answers is unanswered. The buyer was invited to ask and then left waiting.

The two are recorded separately and never merged into a single failure figure. Merging them would penalize a company for an explicit choice and let an unanswered company hide inside the same number.

The measurement window is a forward rule and was not applied to our 2023 data

Read this section before the rule in it. Our own 2023 demo response study applied no observation window of any kind. Collection stopped at two emails per company, not at an elapsed time, and replies in the data arrive as late as 133.6 hours after the form was submitted for a first reply and 149.2 hours for a second email, on a base of 2,386 emails from 1,685 companies. Six first emails and twelve second emails land beyond 96 hours, and every one of them is counted as a late reply and not as a company that never answered.

So nothing below describes how our figures were collected. It is a rule for anyone measuring their own company from now on.

The specification sets a 96 hour window from the start event. A response arriving after 96 hours is recorded as never answered rather than as a long wait.

The reason is arithmetic. A response at 40 days, entered as 57,600 minutes, moves a mean by more than the underlying behavior justifies. Recording it as a category rather than a number keeps the distribution readable.

Never publish the wait without the never-answered share

A censored window discards information, so two figures are always published together and neither is published alone.

1. The median wait among those who answer.
2. The share who never answer.

A median of 30 minutes is a different claim when 90% never reply than when 10% never reply. The median alone describes only the companies that responded.

Our 2023 study publishes the second of those and not the first, and the reason matters. The never-answered share is 93.4% of 1,685 companies, 1,573 of 1,685, 95% confidence interval 90.3% to 96.1%. There is no published median Buyer Wait Time from that study, because the human verdicts are a reweighted estimate across a stratified sample rather than a label on each of the 1,685 companies, so no distribution of human reply times exists across the population.

Two timing figures from the study are published, and neither is a Buyer Wait Time. Across all 1,685 first replies the median wait was 19.7 minutes, a quarter inside 4.1 minutes and a quarter beyond 393.9 minutes. That is time to a first reply of any kind, most of them automated. Among the 190 hand labeled emails the median lag was 765 minutes for the ones a person wrote against 281 minutes for automated acknowledgments, on bases of 17 and 162 verdicts, which is direction rather than a population median.

A company measuring itself under this specification produces a single Buyer Wait Time figure or a never-answered verdict. That is a cleaner measurement than the one our study could make across 1,685 companies, and it is the measurement the specification exists to enable.

How a company measures its own Buyer Wait Time

1. Choose someone outside sales and marketing to run the test, using an email address the company does not recognize. An address on a known domain routes differently and invalidates the result.
2. Submit the form the way a real buyer would, with a plausible company name and a plausible request. Do not flag the submission as a test.
3. Note the exact time of submission, to the minute.
4. Note whether the form actually submitted. Record any error, any validation failure, and any required field the tester could not honestly complete. Form failure is a result, not a setup problem.
5. Tell nobody inside the company that the test is running. A team that knows is measuring its best behavior rather than its normal behavior.
6. Wait 96 hours. Do not intervene and do not follow up. Capture every message that arrives in that period, not the first two.
7. Classify every message received against the classification rules. Record the arrival time of the first response in each category, and record whether any of them was written by a person.

The output is one Buyer Wait Time figure, or a never-answered verdict, plus a record of what the buyer received in the interval.

Step 6 is where this procedure improves on our own 2023 collection. Capturing everything inside a fixed window, rather than stopping at the second message, removes the one limitation that pushes our headline figure in a known direction.

Contrast with Speed to Lead

Speed to Lead measures the interval between a form submission and the seller's first outbound touch, usually as tracked inside the seller's own CRM.

The two metrics differ in subject and in what they count. Speed to Lead makes the company the subject and the buyer an object called a lead. It times the seller's internal handling, and it is generally satisfied by any recorded touch, including automated sequences that no person wrote. A company can post a fast Speed to Lead figure while every buyer in the sample is still waiting for a person.

Buyer Wait Time times the same event from the buyer's side. It asks what the buyer experienced rather than what the system logged. The buyer is the subject, the stop condition is a human response, and automated touches do not count however quickly they fire.

Neither metric substitutes for the other. Speed to Lead is an operations metric for a sales team managing its own process. Buyer Wait Time is a measurement of what the buyer received.

What this specification cannot support

It is a measurement rule, not a benchmark. A single company running this procedure produces one observation. Comparing that observation to our 93.4% is comparing one company to a 2023 sample of 1,685, and the specification supports no significance claim on a base of one.

It cannot be crossed with site type. The study behind the published rates was powered to measure one overall rate and cannot resolve differences between site types, so no Buyer Wait Time figure should be published broken out by form length, phone requirement or chat installation.

It carries no outcome data of any kind. Buyer Wait Time measures an interval. It records nothing about revenue, meetings held, deals won or buyers lost. A short Buyer Wait Time is not evidence of a better outcome, and this specification supports no such claim.

Figures affected by link tracking publish as bounds. Where scheduling destinations are hidden by redirect tracking, as they were across large parts of our 2023 corpus, any share of replies containing a booking link is published as a range or as a floor rather than as a point.

It measures one buyer. A company's Buyer Wait Time on a single test is what one buyer experienced on one day. Repeat the test before treating it as a property of the company.

How to cite this

GTM Clarity, Buyer Wait Time, the metric specification, 2026. Demo response data collected 2023. Website data collected 2021 and 2025. Research directed by Terry Wilson, fieldwork by the GTM Clarity research team.

The report that proposes this metric is The Buyer Wait Time Report 2026.