One thing here was measured. The rest are assumptions, and they are marked.
Most of what this industry publishes as a benchmark is a target somebody thought was reasonable. Some of it is genuinely observed. The two get printed in the same typeface, and a prospective buyer has no way to tell them apart.
On this page they look different. A solid rule with a filled mark means it was measured against a published rubric. A dashed rule with an open mark means it is a planning reference we use when no study exists. That distinction holds for the rest of the page.
- Observed
- Collected and scored against a written rubric, published with its sample, its definitions and a confidence interval on every measure.
- 24 measures, one study
- Planning reference
- A response-time, review-capture or continuity target we apply when scoping work. Reasonable, used in practice, and not a measurement of what comparable businesses achieve.
- 30 profiles
The 2026 Accounting Firm Digital Front Door Benchmark
We collected the public websites of accounting firms across the United States and Canada and scored what a prospective client could actually establish from them: whether services were explicit, whether accountable people were named, whether there was a route for a new client as distinct from an existing one, whether a booking path explained what the meeting was for, and whether anyone was told what happens after they make contact.
Contactability was not the problem. 86% of sampled sites exposed a recognisable place to begin contact. What was rare was guidance: an explicit new-client route, an explanation of the booking, a stated response expectation, or any indication of what happens outside office hours.
- Candidates frozen before collection
- 240
- Sites reachable
- 160
- Eligible and scored
- 118
- United States / Canada
- 85 / 33
- Pages observed
- 816
- Measures published
- 24each with a 95% interval
- Rubric / method version
- rubric.v1.1 · afdfb-2026-v1
Published August 2026. Convenience sample; not nationally representative. The complete dataset is published as JSON and CSV.
Selected measures, with their intervals.
Eleven of the 24 published measures, ordered from what the sample did well to what it did not. The pale band is the 95% interval. The dark tick is the point estimate. Read the band first.
- Dedicated contact destinationNavigation or page content exposes a dedicated contact route.86.4%102 of 118 · CI 79.1–91.5
- Consultation path is visibleA visible consultation, appointment, or scheduling path exists; it is not activated.83.1%98 of 118 · CI 75.3–88.8
- A human route is explicitA visible route reaches a named person, office team, or staff-assisted channel.59.3%70 of 118 · CI 50.3–67.8
- Office or availability hours are visibleHours or a clear availability window appear publicly.33.9%40 of 118 · CI 26–42.8
- Secure document route is visibleA secure upload or document portal route is visibly identified without entering it.18.6%22 of 118 · CI 12.6–26.6
- New-client route is explicitA visible label or instruction identifies a route for prospective or new clients.16.9%20 of 118 · CI 11.2–24.7
- Response expectation is visibleVisible copy states when or how a prospect should expect a reply.14.4%17 of 118 · CI 9.2–21.9
- Booking next step is explainedThe page indicates what the meeting is for, who it is with, timing, or what follows.13.6%16 of 118 · CI 8.5–20.9
- New-client intake is visibleA public route or explanation identifies how a new client begins intake.11%13 of 118 · CI 6.6–17.9
- After-hours expectation is clearThe site explains after-hours handling, next-business-day response, or emergency limitations.1.7%2 of 118 · CI 0.5–6
- Sensitive-document boundary is explainedVisible copy tells prospects where sensitive documents should or should not be sent.0%0 of 118 · CI 0–3.2
2 measures
The sample was reliably contactable
Most sampled sites gave a visitor a recognisable place to begin contact, and a consultation-oriented next step was usually visible. Being reachable was not the problem.
3 measures
Being reachable is not the same as being guided
Four in ten sampled sites did not make a human-assisted route explicit under the rubric. Two thirds did not publish hours or an availability window. Most did not visibly identify a secure document route on the pages observed.
6 measures
What happens next was rarely explained
This is the cluster worth sitting with. A visitor often had to infer whether a contact route was intended for new clients at all. Most sampled sites did not set a public expectation for what happens after contact, and the presence of a booking path did not usually explain the experience that followed. After-hours guidance was close to absent, and no sampled site told prospects where sensitive material should or should not be sent.
Percentages describe the 118 eligible sampled sites under the definitions shown, and nothing beyond them.
Where the public journey stopped.
Four descriptive stages synthesised from seven directly observed signals. They describe how far a visitor could get, not how good a firm is. Mutually exclusive descriptive stages derived from seven directly observed journey signals; not a quality score or national estimate.
- No clear contact path97.6%
- Contactable only6151.7%
- Partly guided4538.1%
- Connected journey32.5%
Three of 118 sampled sites reached the connected stage. That is the finding worth sitting with: the visible gap is not design quality, it is the absence of a guided path from a stranger’s first question to a prepared first conversation.
Method, and what it refuses to claim.
The limits are published with the findings rather than after them. A study that only reports what it found is advertising.
What counted as the sample
240 candidate firms were frozen before collection began. 160 sites were reachable and 118 met the eligibility rule. 85 were in the United States and 33 in Canada. 816 pages were observed in total.
What the sample is not
A convenience sample of public websites. It supports statements about what the sampled sites showed. It does not support a national estimate, a ranking of firms, or a claim about firms outside the sample.
What was never entered
Forms were observed and never submitted. Portals and secure upload routes were identified from the public interface and never entered. No firm names or firm-level results are published.
What was deliberately excluded
The rubric refuses judgements it cannot make reliably from public pages: differentiation quality, value proposition quality, mobile experience quality. Those are absent from the findings rather than estimated.
The practitioner layer is labelled separately, and deliberately
The white paper also carries recurring patterns from more than 100 conversations with accounting-firm principals. Those observations were not coded for prevalence and they establish no causal relationship, so they carry no percentage anywhere on the site.
The published boundary reads: “Recurring practitioner patterns only. No prevalence estimate, benchmark percentage, or causal claim is permitted.”
Signals the descriptive stages were built from
- public contactability
- prospect routing
- booking explanation
- intake visibility
- response expectation
- human route
Planning references.
30 operating profiles carrying the response-time and review-capture targets we apply when scoping work in a sector. Every one of them is an assumption. None of them was measured, and each carries the open mark for that reason.
- Home Services BenchmarkField Response BenchmarkFirst response assumed under 8 min
- Restoration BenchmarkFirst-Notice BenchmarkFirst response assumed under 6 min
- Basement Waterproofing BenchmarkInspection Confidence BenchmarkFirst response assumed under 15 min
- Garage Door Repair BenchmarkUrgent Dispatch BenchmarkFirst response assumed under 5 min
- Dental BenchmarkPatient Intake BenchmarkFirst response assumed under 15 min
- Personal Injury BenchmarkCase Intake BenchmarkFirst response assumed under 5 min
- Med Spa BenchmarkConsult Conversion BenchmarkFirst response assumed under 12 min
- Plastic Surgery BenchmarkConsult Trust BenchmarkFirst response assumed under 30 min
- Accounting & Advisory BenchmarkConsult Funnel BenchmarkFirst response assumed under 45 min
- Urgent Care BenchmarkVisit Capture BenchmarkFirst response assumed under 10 min
- Home Automation & AV BenchmarkPremium Project BenchmarkFirst response assumed under 30 min
- Medical Specialists BenchmarkReferral Confidence BenchmarkFirst response assumed under 20 min
- Premium Concierge BenchmarkLuxury Inquiry BenchmarkFirst response assumed under 15 min
- Wealth Management BenchmarkAdvisory Trust BenchmarkFirst response assumed under 45 min
- Commercial Insurance BenchmarkRenewal Authority BenchmarkFirst response assumed under 45 min
- Veterinary BenchmarkHousehold Trust BenchmarkFirst response assumed under 8 min
- Funeral Home BenchmarkFirst-Call Trust BenchmarkFirst response assumed under 5 min
- Family Law BenchmarkRetainer Momentum BenchmarkFirst response assumed under 10 min
- Elder Law & Estate Planning BenchmarkFamily Decision BenchmarkFirst response assumed under 15 min
- Appliance Repair BenchmarkSame-Day Service BenchmarkFirst response assumed under 12 min
- Junk & Moving BenchmarkEstimate & ETA BenchmarkFirst response assumed under 15 min
- Tree Service BenchmarkStorm Intake BenchmarkFirst response assumed under 8 min
- Solar Installation BenchmarkConsult Qualification BenchmarkFirst response assumed under 20 min
- Auto Glass BenchmarkInsurance Appointment BenchmarkFirst response assumed under 12 min
- Pool & Spa BenchmarkSeasonal Demand BenchmarkFirst response assumed under 30 min
- Design-Build & Remodel BenchmarkProject-Fit BenchmarkFirst response assumed under 60 min
- Landscape Design-Build BenchmarkDesign-Budget BenchmarkFirst response assumed under 180 min
- Locksmith Emergency BenchmarkAfter-Hours Dispatch BenchmarkFirst response assumed under 5 min
- Fire Safety & Security BenchmarkDispatch & Inspection BenchmarkFirst response assumed under 15 min
- Collision Repair BenchmarkInsurance Approval BenchmarkFirst response assumed under 30 min
The response target shown is the workable threshold from each profile. Each page also carries a stricter and a fragile threshold, a review-capture assumption, and the reasoning behind them. Treat all of it as a starting position for a conversation about pace.
Using any of this in a decision.
Four rules, which we also apply to ourselves.
Use an observed measure as a description of the sample.
A finding on this page means that many of the 118 collected sites showed a thing under a written definition, within the interval published beside it. It is not a statement about the profession, and the interval is part of the finding rather than a footnote to it.
Use a planning reference as a target you are free to argue with.
A response-time or review-capture target is a working assumption we apply when scoping. It is a reasonable place to start a conversation about pace. It is not a measurement of what comparable businesses achieve.
Never present a planning reference as industry performance.
This is the single failure this page is designed to prevent. If a number here is not carrying the observed mark, it was not measured, and it should never appear in a deck as though it were.
Check the interval before the headline.
A measure at two percent with an interval reaching six percent, and a measure at eighty-six percent with an interval reaching ninety-two, are different kinds of statement. The plot on this page draws the interval first and the point estimate second for that reason.