Who answers?

The headline

34 of 84 (40%) never responded by any channel. Not in 72 hours, and for most of them not at all; the longest tracked silences ran past 180 hours. Both after-hours calls went unanswered with no callback.

Of the 50 that did respond, the median time to a human was 12.6 hours. Eight answered within 15 minutes. Fifteen answered within an hour. Thirteen took more than a day, and six took more than two.

So the market is not slow. It is split. Of every ten businesses, one answers a new customer within fifteen minutes, five answer later, mostly the next business day or after, and four never do.

When you ask matters more than who you ask

Inquiries went in during four windows. Different businesses in each, so this is a pattern, not a controlled test.

Friday evening is contested ground: a real share of shops are still working at 6pm, and half of the ones who answered did so within the hour. The other half waited for Saturday or Monday, and that produced the three longest replies in the study, 63 to 69 hours. Sunday evening had the highest silence rate, and every Sunday reply arrived Monday morning between 7:27 and 12:55. Not one Sunday inquiry was worked that night.

By trade

Plumbing is the clearest example of the split: more than half never answered, and the ones who did were the fastest in the study. Pest control answered more often but slowest, mostly by phone two days later.

An auto-reply is not a reply

16 businesses acknowledged the inquiry automatically, most within a minute. What happened after the acknowledgment:

• 5 had a person follow up within about two hours.
• 4 had a person follow up between 12 and 45 hours later.
• 7 never had a person follow up at all.

Some of the automated messages made promises. One said “someone will be reaching out shortly,” 42 seconds after the form; the human arrived 44.7 hours later. One confirmation screen said “we will contact you within the next 20 minutes”; nobody did. One shop sent three scheduled texts over four days and no human message. One “request an estimate” form’s only response was a mailing-list opt-in email. One text-back quoted my message back, asked if I was a current customer, and when I said no, replied that they could not respond by text and I should call the office.

Published promises do not predict behavior

34 of the 84 businesses publish a response promise on their site: 24/7, same-day, emergency service, or a specific number of minutes. Those 34 were slightly more likely to go silent than the 50 without a promise: 16 of 34 (47%) never responded, against 18 of 50 (36%).

The tightest promises missed by the most. A site that promised contact “within 15 minutes” made one call 12.5 hours later and left no voicemail. A site that lists 7am to 8pm, seven days, took a Friday 6pm inquiry inside its own posted hours and never answered it.

Ten weeks later, the same businesses

Eleven businesses were shopped twice, on June 8 and again on August 16 or 29, with a different inquiry each time.

• 5 were silent both times.
• 3 got slower (20.5 to 48.1 hours; 17.2 to 34.0; 18.9 to 21.2).
• 1 answered in June (16.8 hours) and was silent in August.
• 1 was silent in June and answered in August, at 13.9 hours.
• 1 went from 17.1 hours to 46 minutes after installing a text-back tool between the two visits.

Response behavior looks like a trait of the business rather than a bad week. The one business that changed materially did it by changing its system, not its habits.

The phone matters, for them and for the measurement

Trades owners call. In the August 29 session, five of the thirteen businesses that responded did so only by phone, voicemail or text and never sent an email. Several others replied from a personal Gmail rather than the address on their site. A measurement that reads only the inbox would have scored all of them silent.

This study checked email, text and voicemail for every zero. It did not always: an automated pass on September 1 scored twelve businesses silent from the inbox alone, and six of them were flipped the same afternoon, five from the callback line’s log and one from an email sent from a personal address that the matching had missed. Those corrections are in these numbers. It is the reason every zero here is gated on three channels, and the reason I would not trust a response study that cannot say how it handled the phone.

Fast is not the same as answered

From August 22 on, every inquiry asked what the visit would cost, and I tracked whether the first human reply gave a number. Of the human replies to that question, about half gave a price or a bounded range (“$245 for a conventional cable, two-month warranty”; “$150 plus parts for roughly the first hour”; “$149 diagnostic, Thursday 10 to 12 is open”). The other half asked me to call, sent a booking link, or asked for photos.

One shop’s automated text arrived the same minute as the form and asked a good diagnostic question. Getting a price out of the same shop took three asks over 47 hours. The reply was two characters: “$140.”

What one location looks like on a scorecard

This is the row format used for every business in the study, shown here for six real rows with names removed.

Across a portfolio, the same table becomes a league table by location, a trend per location per round, and a short fix list: which forms are broken, which auto-replies promise something nobody delivers, which locations answer by phone and never by text, and which windows are uncovered.

Method and limits

One persona, one inquiry per business per visit, one window per visit. The 72-hour cutoff is a floor for the silence count, not a ceiling for the delay; most silent businesses stayed silent well past it. Businesses were selected, not sampled at random: owner-operated (46 of 84 confirmed by name), roughly 10 to 150 Google reviews (median 63), with a working form or published email. National franchises, businesses without a website, forms that required a street address, and booking-only sites were excluded. Scores were assigned against a written rubric, with the reasoning for each score and the scores next to it recorded in the row. Timestamps are from the inquiry inbox, the callback line’s call and text log, and the auto-reply headers. Quotations are lightly paraphrased so no business can be identified by search. The June round (14 businesses, 6 silent) used the same method and is included only in the repeat-visit section.

About

I am Blake Rosenthal, in Newbury Park. I built and ran a lead-response system for a medical clinic in Thousand Oaks, measured its first 30 days (median 2 minutes from web inquiry to booked consultation, 21 consultations), and published the numbers with their limits. This study is what I learned by turning the same measurement on the businesses around me.

Questions about the data, or about running this across a set of locations: blake@operaiteconsulting.com, (805) 372-2528.