Tri-State Accounting Business Study
Search visibility, websites and AI descriptions of 1,973 accounting-related firms in New York, New Jersey and Connecticut. Sampling: August 13 to October 15, 2026.
On this page
Sampling period: August 13 to October 15, 2026.
Read the full report (PDF) · Download the report, tables and figures (ZIP)
What the study found
Yelp held 93 first-page positions across the accounting searches in this study. No individual firm in the sample held more than 19. The comparison shows how widely a directory can appear across a regional search, while an individual practice may serve a much narrower area.
Across 93 completed Google searches, firms in the study occupied 206 of 830 first-page organic positions, or 24.8 percent. Directories, review and lead-generation platforms, social sites and national tax-preparation chains held at least 258 positions, or 31.1 percent. Other positions went to business, professional, government and community websites.
The study examined 1,973 accounting-related firms. Of these, 1,296 were listed in New York State, including 593 in New York City; 487 were listed in New Jersey and 190 in Connecticut. These are locations reported in a commercial business database, rather than a census of licensed accounting practices.
Larger firms appeared more often in the completed organic searches. The observed rate was 1.7 percent among firms reporting 1 to 10 employees, compared with 11.0 percent among those reporting 50 or more. This describes the sample; it does not show that firm size caused greater visibility.
In a separate analysis, three AI systems described sources of information about 546 named firms. Of 5,049 source descriptions, 471, or 9.3 percent, referred to a firm's own website or content. The remainder referred to third-party sources. These are the systems' descriptions of their sources, not independently verified citations or recommendations.
The firms studied
The starting list contained 2,009 accounting-related business records. Applying the study's selection criteria left 1,973 firms for analysis; 30 records were excluded and six could not be resolved.
| Listed location | Firms | Share of the sample |
|---|---|---|
| New York State | 1,296 | 65.7% |
| New York City, included in New York State | 593 | 30.1% |
| New Jersey | 487 | 24.7% |
| Connecticut | 190 | 9.6% |
| Total, without counting New York City twice | 1,973 | 100.0% |
The list was not randomly selected. Business descriptions, locations and employee counts came from the data provider and were not independently checked against licensing registers. The findings describe this sample, not every accounting firm in the three states.
How the study was conducted
The sampling period was August 13 to October 15, 2026. The study combined three separate forms of research: searches for accounting services, automated checks of firms' websites and domain settings, and questions to AI systems about firms that had already been named.
The organic search design used five accounting-service phrases across 19 cities and areas, for 95 planned searches. Ninety-three returned usable results. The analysis counted 830 first-page organic result positions across 205 distinct websites. Advertising and the separate local-search results were not included in those 830 positions.
Each result was compared with the websites of the firms in the sample. Other websites were grouped by purpose, including directories and review platforms, social sites, national tax chains, professional or public bodies, and other business websites. Uncertain classifications were placed in the other-business category, making the reported platform-and-chain count conservative.
Automated website and domain checks recorded what could be observed at the time of the check. A separate local-search analysis recorded listings and their positions. The AI analysis asked systems from Gemini, OpenAI and Perplexity about named firms and recorded how they described their information sources.
Results from these different methods are reported separately. They do not measure customers' decisions, website traffic, revenue or the quality of professional services.
Who appeared on the first page
| First-page result holder | Positions | Share of 830 |
|---|---|---|
| Firms in the study | 206 | 24.8% |
| Directories, lead-generation and review platforms | 201 | 24.2% |
| Social and community platforms | 20 | 2.4% |
| National tax-preparation chains and software | 37 | 4.5% |
| Professional bodies, government, education, media and civic sites | 44 | 5.3% |
| Firm and business websites outside the sample | 322 | 38.8% |
Yelp accounted for 93 positions. The most visible individual firm in the sample accounted for 19. These are appearances across the complete set of regional searches; a website could appear more than once.
The 322 positions held by business websites outside the sample show why the sample should not be confused with the whole market. Those businesses were not classified as local or nonlocal, and their exclusion from the sample does not establish why they were not included.
Directories can help prospective clients compare providers. Their presence is not evidence that an accounting firm is losing customers. The practical finding is that a business's own website is only one of the places where a client may first encounter it.

Firm size and visibility
| Reported employees | Firms | Found in organic results | Found in local-search results |
|---|---|---|---|
| 1 to 10 | 1,409 | 24 (1.7%) | 56 (4.0%) |
| 11 to 49 | 443 | 16 (3.6%) | 28 (6.3%) |
| 50 or more | 109 | 12 (11.0%) | 11 (10.1%) |
Twelve firms with a missing or zero employee count were excluded from this comparison. Employee counts were reported by the data provider. The observed differences do not establish cause and effect.
DMARC is a domain record that sets an email-authentication policy. Its presence also varied by reported firm size: 531 of 1,291 checked small-firm domains had a record (41.1 percent), compared with 239 of 422 in the middle group (56.6 percent) and 74 of 99 in the largest group (74.7 percent). A record's presence alone does not establish that it is effective or correctly configured.

Website and domain checks
The automated website checks recorded one final outcome for each of the 1,973 firms.
| Recorded outcome | Firms | Share |
|---|---|---|
| Website reachable | 1,364 | 69.1% |
| Website unreachable | 401 | 20.3% |
| Check timed out | 148 | 7.5% |
| Check was blocked | 33 | 1.7% |
| Website returned an error | 27 | 1.4% |
An unsuccessful automated check does not prove that a website was unavailable to a human visitor or explain the cause. The results should not be used to judge a firm's service quality.
Domain information could be checked for 1,823 firms. The other 150 remain unknown for the measures below and are excluded from their percentages.
| Domain setting, among 1,823 checked | Observed | Not observed |
|---|---|---|
| DMARC email-authentication policy | 848 (46.5%) | 975 (53.5%) |
| SPF authorized-sender record | 1,535 (84.2%) | 288 (15.8%) |
| Content Security Policy website header | 246 (13.5%) | 1,577 (86.5%) |
These checks recorded presence, not configuration quality. They were not security audits and do not establish whether a firm was vulnerable to abuse.
Search coverage and missing results
The organic searches completed for 93 of the 95 planned queries. The two unsuccessful searches were "tax accountant in The Bronx" and "CPA firm in New Haven."
At least 53 firms in the sample appeared in the completed first-page results. Of these, 29 had a best observed position in the top three and 24 had a best position between fourth and tenth. The remaining 1,920 were not matched in the completed searches. Because two searches failed, they cannot be described as absent from the full planned search set.
The separate local-search analysis completed all 95 searches. It matched 95 firms: 22 with a best recorded position in the top three and 73 between fourth and twentieth. The other 1,878 were not matched in those searches. This is not a finding that they were absent from local search generally.
A firm can appear in both organic and local-search results, so these counts must not be added together. Search results vary by location, language, device, wording and time. The study used localized English desktop searches and describes only that selected set.
How AI systems described their sources
Complete results from Gemini, OpenAI and Perplexity were available for 546 firms, or 27.7 percent of the sample. The analysis asked about firms that had already been named; it did not test which accountant an AI system would recommend to a prospective client.
| Source described by the AI system | Mentions | Share of 5,049 |
|---|---|---|
| The firm's own website or content | 471 | 9.3% |
| Third-party sources, including directories, reviews and profiles | 4,578 | 90.7% |
Gemini accounted for 1,414 recorded source descriptions, OpenAI for 1,588 and Perplexity for 2,047. The study recorded descriptions rather than page addresses, so these entries cannot be used to verify the specific pages consulted. Nor do they establish whether the information was accurate.
The AI findings cannot be generalized to the 1,427 firms without complete results. No firm-by-firm comparison or measure of recommendation frequency is reported.

What the findings mean
An accounting firm's online presence is distributed across its own website and third-party pages. In the searches examined here, directories and other platforms were prominent, and the separate AI analysis frequently recorded descriptions of third-party sources.
For firm owners, that makes the accuracy of services, qualifications and contact details across those pages worth checking. For prospective clients, directories and AI summaries can be starting points for comparison, while credentials, scope of work and suitability still require direct verification.
This study cannot establish that changing a website or a domain setting would produce more customers. It did not measure buying decisions, leads or financial results. It also draws no conclusions about publishing frequency or content quality because the available observations were incomplete.
About the study
Markster produced and funded this research. The company develops and operates business software and provides specialist business services. Its commercial work includes business visibility and operations.
The report publishes combined findings without identifying or assessing individual accounting firms. Percentages are rounded to one decimal place. The downloadable tables provide the counts and denominators used in the report.
For questions or corrections, contact press@markster.ai. Cite the study as: Markster, Tri-State Accounting Business Study, published November 21, 2026; sampling period August 13 to October 15, 2026.
Report and figures
The download includes the full report, study tables, methods and limitations, and four figures. No signup is required.
- Full report (PDF)
- Report, tables, methods and figures (ZIP)
- First-page search positions (PNG)
- Yelp comparison (PNG)
- Visibility by firm size (PNG)
- AI source descriptions (PNG)
Updated November 21, 2026: the sampling period has been corrected to August 13 to October 15, 2026.