We submitted a real inquiry to 273 US owner-operated service businesses across commercial cleaning, dental and aesthetics, family law, mortgage and lending, and trades. We recorded whether the business had a working web inquiry form, whether a reply arrived, how long it took, on which channel, and whether it came from a person or a system.
Two findings.
Summary
One in five businesses had no web inquiry form
56 of 273, or 20.5%, had no form present, a social link substituted for one, or a submission container that did not function. In dental and aesthetics the figure is 42.9%. This measures web forms only and does not assess direct email links (`mailto:`) or telephone contact.
Seven in ten delivered inquiries were never answered
Of 211 inquiries provably delivered, 29.9% received any reply within 72 hours and 21.3% received a reply from a person. 70.1% received nothing at all.
Existing figures in this area trace back to research published in 2011 and 2012, conducted on B2B sales pipelines. This study is primary, current, and measures owner-operated service businesses directly.
Why this study exists
Ask what proportion of businesses respond to an inquiry and the answers conflict. Published figures range from under 30% to over 60% contacted, and almost none state a denominator or a censoring window. Nearly all trace to the same two sources: a Harvard Business Review study of B2B lead pipelines and an InsideSales response-management study, both over a decade old.
Nobody has measured whether a small US service business can receive a web inquiry at all. That is the question a customer actually faces.
Method
Sample
273 US owner-operated service businesses, drawn across five sectors.
The inquiry
One base message per sector, held constant in persona, tone, ask and sign-off. Family law and mortgage were sent verbatim with no variation. Cleaning, trades and dental substituted the named service according to a substitution table written before fieldwork began, so a plumber received a water heater inquiry and an electrician received an electrical one.
The inquiry is standardized, not identical. A single message cannot be plausible to both a divorce lawyer and a roofer, and an implausible inquiry measures implausibility rather than responsiveness.
- Commercial cleaning — “Hi, I manage a small office, around 2,000 sq ft, and we're looking for regular weekly cleaning. Can you send a quote or let me know what you need from me? Thanks”
- Dental & aesthetic — “Hi, I'm new to the area and looking into [treatment]. Can you let me know availability and rough pricing for a first visit? Thanks”
- Family law — “Hi, my partner and I are looking at an amicable separation and I'm trying to understand our options before we decide anything. Could someone let me know how you usually handle this? No rush, thanks”
- Mortgage & lending — “Hi, I'm a first-time buyer just starting to look at options and wanted to understand your process and current rates. What's the best way to get started? Thanks”
- Trades — “Hi, my water heater is getting old and I'm thinking about replacing it in the next month or two. Can you give me a rough quote or let me know what you'd need to see first? Thanks”
Delivery proof
An inquiry counts as delivered only on positive evidence. A submission is stamped delivered when it produces a redirect to a success path, confirmation text inside the submitted form container, or the submitted fields returning empty. Absent all three, the row is not counted as sent.
The default is no proof, no row. A submission that appears to work but produces no evidence is treated as not delivered, not as delivered-and-ignored. This is deliberately conservative: it removes rows from the denominator rather than inflating the non-response rate.
A later reply is itself proof of delivery, since a reply cannot exist without the inquiry arriving.
Censoring
A business is counted as a non-responder only after 72 hours have elapsed from its own submission time. The window is computed live per row, never against a fixed calendar date, so every business receives the same observation period regardless of when it was contacted.
Rows still inside their first 72 hours are not counted either way. They are unfinished observations, not silence.
Classification
The rule below was written on August 5, 2026, before the classification pass it governs, and applied unchanged.
- Reply arrives within 5 minutes of submission — automated.
- Reply is one of two or more near-identical messages in the same thread — automated.
- Reply engages something specific to the actual submission — human.
- Otherwise, tie-break on signature: a named individual is human, a generic department address is automated. Every tie-break is flagged as lower confidence.
Definitions
- Contacted — A business in the sample, whether or not an inquiry could be submitted.
- No web inquiry form — No form present, a social link substituted for a form, no fillable form found, or a form confirmed non-functional. Assessed against forms only. Direct email links (`mailto:`), telephone numbers, and click-to-call were not assessed and do not count as a form.
- Delivered — Submission met the delivery-proof rule above.
- Response — Any reply of any kind, on any channel, within 72 hours.
- Substantive response — A reply classified as human under the rule above.
Finding 1 · One in five businesses had no web inquiry form
Measured at the point of contact, across all 273 businesses.
56 of 273 businesses, 20.5%, had no working web inquiry form — no form present, a social link substituted for one, or a submission container that did not function.
This measures the presence and functionality of web forms only. It does not assess direct email links (`mailto:`), telephone numbers, or click-to-call, and some businesses in this group are reachable by those routes. The finding is that a written web inquiry, the route most customers reach for first, had nowhere to land.
This is a property of the business, not of the study instrument. It is measured before any inquiry is sent and does not depend on response timing, censoring, or classification.
Dental and aesthetic practices are the outlier. More than two in five had no route by which a prospective patient could submit an inquiry online.
Finding 2 · Seven in ten delivered inquiries are never answered
Of 273 businesses contacted, 217 inquiries were submitted and 211 were provably delivered. Six are excluded because delivery could not be established either way.
By sector
Mortgage and lending is the weakest sector by a wide margin. Of 40 delivered inquiries, six drew any reply and exactly one came from a person. The other five were automated acknowledgments.
Dental and aesthetics is weak on both measures. It has the highest rate of missing inquiry forms at 42.9%, and among practices that could receive an inquiry, 22.6% replied. A prospective patient faces two failure points rather than one.
How long a reply takes
Median time to first response: 5.1 hours, across 58 timed responses.
Ten replies arrived within five minutes. Almost all were automated acknowledgments rather than a person, and under the classification rule a reply inside five minutes is recorded as automated.
Counting only replies from a person, the median rises to 8.8 hours. The fastest was 7 minutes and the slowest 56 hours.
Two replies arrived after the 72-hour window closed and are shown above for completeness. One was an automated missed-call text sent 16 days later by a phone system, which is reported in the data but excluded from any claim about how long businesses take to reply.
When the inquiry was sent
Each submission was assessed against the receiving business's own local opening hours.
An inquiry arriving outside opening hours was answered at a little over half the rate of one arriving during them, measured over the same 72-hour window in both cases.
This is the finding that matters most to a customer, because a customer does not wait for opening hours. Evenings and weekends are when people research a service, and they are when an inquiry is least likely to reach anyone.
Opening hours were recorded for 167 of the 211 delivered inquiries. The remaining 44 are not included in this comparison.
Key takeaways
Every figure below is from this study. No external benchmark is used.
- A written web inquiry has nowhere to land at one business in five. 20.5% had no working form. Those customers never reach the queue, so no amount of follow-up discipline recovers them.
- Seven in ten delivered inquiries get nothing back. 148 of 211 received no reply of any kind within 72 hours.
- A reply from a person is rarer still. 21.3% of delivered inquiries were answered by a human being.
- The median reply takes 5.1 hours, and 8.8 hours if it comes from a person. Half of all people who did reply took longer than a working day to do it.
- Fast replies are mostly machines. Ten replies arrived within five minutes. Almost all were automated acknowledgments.
- Out-of-hours inquiries go unanswered. 19.8% against 34.8% for inquiries arriving during opening hours, on the same 72-hour window. Customers research in the evening. Businesses respond in the morning, if at all.
- Some sectors are much worse than others. Mortgage and lending answered 15.0% of delivered inquiries and only 2.5% with a human reply. Dental and aesthetics combines the worst form coverage with a low response rate.
- The gap is capture, not generation. Every business in this sample already had demand arriving. What varied was whether anything happened to it.
What this means for a business spending on demand
A business paying for inquiries is paying for the top of a funnel whose entry point may not work and whose response may not come. This study measures how often that is true. It does not measure what it costs, because that depends on the business.
Rosebud Solutions operates the layer between an inquiry arriving and a booking being made: capture and response on every channel, qualification against the business's own rules, booking into a live calendar, reminders, and reactivation of leads that went cold. It is operated by Rosebud rather than configured by the client, and it stops at the booking, where the client's team takes over.
The study was run on the same platform Rosebud operates for that work, which is why the question was worth asking and why the method is published in full above. Pricing and plans · How it works
Limitations
- Sector pools are small. Delivered inquiries per sector range from 31 to 47. Sector rates carry a wide confidence interval and the ordering between adjacent sectors should not be treated as established. The gap between the highest and lowest sectors is large enough to be meaningful; the gap between neighbours is not.
- Sends were concentrated on five days. July 26, July 28, August 4, August 7 and August 10. Day-of-week effects are not separable from batch effects, which is why the timing comparison is reported against each business's own opening hours rather than by calendar day.
- Opening hours were recorded for 167 of 211 delivered inquiries. The out-of-hours comparison uses that subset.
- Five responses could not be timed and are excluded from the time-to-response figures while remaining in the response counts.
- Six businesses are excluded because delivery could not be established either way. Four had a manual resubmission that could not be independently confirmed. Two returned a blank form with no error and no confirmation. All six are excluded rather than assumed in either direction.
- Response is not conversion. This study stops at first response. It does not measure whether a reply led to a booking, a quote, or a customer.
- Single inquiry per business. Each business received one inquiry. A business that missed one inquiry might answer the next, and a single test cannot separate a systemic failure from an owner on holiday or a short-staffed week.
- Spam filtering is unobservable. A form submission that reaches the business but is filtered before a person sees it is indistinguishable, from outside, from one that was seen and ignored. Both are recorded as no response. This applies to all form-submission research and cannot be measured from the sending side.
- Alternative contact routes were not assessed. Finding 1 measures the presence and functionality of web forms. Direct email links (`mailto:`), telephone numbers, and click-to-call were outside scope. A business counted as having no form may still be reachable by other means.
Data availability
Aggregate figures are published here in full and as a machine-readable file: response-study-2026-aggregate.csv, covering every metric in this report broken out by sector.
Per-business records are held internally and are not published, since the study identifies named businesses that did not respond.
Method, definitions, classification rule and exclusion criteria are published above in full so the figures can be independently assessed.
Sector findings
Each sector's figures are set out on the industry page for that sector. Insurance and real estate were not in the study and carry no figure.
Citation
Rosebud Solutions (2026). The 2026 US Service Business Response Study: what happens when you send a US service business a web inquiry. Rosebud Global Ltd. Available at https://rosebud.global/resources/2026-us-service-business-response-study
Published under CC BY 4.0. Reuse is permitted with attribution.
Fieldwork July 26 to August 10, 2026. Response logging closed August 13, 2026. All figures final.
Rosebud Solutions is a Rosebud Global Ltd company (Co. No. 16623472). Rosebud Solutions operates inquiry handling for service businesses, which is why this question was worth measuring. The study was self-funded and no business in the sample is a client.
