We reviewed 200 insurance agency websites, with four websites from each of the 50 US states, in September 2026. Alaska and Hawaii have the same representation as California, Texas and New York. Each website appears once in the sample.
The audit examined the source of one directory-linked page per website for contact links, quote wording and technical settings. Telephone links appeared on 124 pages, quote wording in links or buttons on 132, and canonical URL elements on 180. These findings give agency owners specific places to begin a website review.
Download the complete dataset or check the state-by-state counts.
What we found across 200 websites
| Finding | Pages | Share of sample |
|---|---|---|
| Telephone link | 124 of 200 | 62% |
| Quote wording in a link or button | 132 of 200 | 66% |
| Email link | 84 of 200 | 42% |
| HTML form element | 110 of 200 | 55% |
| Canonical URL element | 180 of 200 | 90% |
| At least one parseable JSON-LD block | 144 of 200 | 72% |
All percentages describe this sample. A source check establishes that an element was detected; testing the customer’s actual task establishes whether it works. Script-rendered content and features on other pages may not appear in these observations.
Make the phone number usable on a phone
Contact-related elements detected in page source
A telephone link lets a visitor tap a number to start a call. We detected one on 124 of the 200 inspected pages. On the other 76, the next useful check is whether the rendered page offers a working call action, displays an unlinked number or directs visitors to a separate contact page.
For your own agency, open the site on a phone and follow the path a new customer would take. Find the office number without searching through several pages. Tap it and confirm that the phone proposes the correct number. Repeat this on your main service and location pages, especially if those pages receive visitors directly from search or advertising.
A multi-office site should make office selection clear before a call starts. A prominent number is less useful if the visitor cannot tell which team will answer. The multi-location SEO guide explains how to connect office pages with their contact details.
Check where the quote action leads
Quote wording appeared in a link or button on 132 pages, or 66% of the sample. The wording alone does not tell a visitor whether they will receive an online price, request a callback or enter an external quoting system.
Match the label to the next step. If the agency reviews a request before contacting the customer, “Request a quote” is clearer than promising an instant result. If a button opens a carrier or comparative quoting tool, explain the transition and offer a way to contact the agency for help.
Walk through the route for each major insurance line. Check that the destination supports that line, works on a phone and identifies the agency. A home insurance button that opens an unrelated general form creates an avoidable decision for the visitor.
Use the conversion review worksheet to record the entry page, intended action and destination before changing the layout.
Test the inquiry route, not the form count
The audit detected an HTML form on 110 pages and an email link on 84. Neither number is a measure of inquiry quality. A form element can belong to search or newsletter signup; an agency can also offer a useful contact route on another page or through an embedded tool.
Choose the contact methods your team can reliably handle. For a form, ask only for the information needed to route and respond to the inquiry. Explain what happens after submission. Keep the confirmation visible and provide a recovery path if submission fails.
Then send a controlled test request and confirm that it reaches the intended inbox or system. For multiple offices or service categories, test each routing rule. Repeat this after changing the form, notification settings or connected software. An attractive form that sends requests to an unattended destination still needs repair.
Verify technical settings beyond their presence
Technical elements detected in page source
All 200 collected pages used HTTPS. Viewport metadata appeared on 196, canonical URL elements on 180 and at least one parseable JSON-LD structured-data block on 144.
These settings answer different questions. HTTPS concerns the connection. Viewport metadata helps a browser size the page for a device. A canonical element identifies a preferred page address. Structured data can describe the business and page in a machine-readable format.
Ask the person maintaining your site to verify the values and behavior:
| Check | What to verify |
|---|---|
| Secure connection | Important pages open without certificate or mixed-content warnings |
| Mobile layout | Navigation, text and form controls remain usable on a small screen and with larger text |
| Canonical URL | Each important page points to the intended public URL, including the correct domain |
| Structured data | Business names, office details and URLs match the visible content |
A detected setting can still contain the wrong value. Conversely, a missing structured-data block does not establish that a page cannot appear in search. Use the insurance agency SEO guide for the broader search review and the website speed guide for loading and interaction checks.
Turn the findings into a short repair plan
Start with one customer task: contacting the correct office about a service the agency offers. Record the page, the obstacle, the person responsible and a concrete acceptance check.
- Repair failed contact paths. Correct wrong numbers, broken destinations and requests that do not reach the team.
- Clarify the next step. Make quote labels, office selection and submission confirmations easy to understand.
- Correct technical settings. Check public URLs, mobile behavior and business information after a repair or migration.
Use the audit worksheet to track those checks. A focused repair may resolve the problem; the redesign checklist helps assess when the underlying structure needs more work.
How the 50-state sample was built
We collected the observations on September 23, 2026 using public listings in the Travelers independent-agent directory. Candidates were placed in a fixed order within each state. We selected four eligible websites per state, excluded duplicate website hosts across the sample and replaced ineligible destinations with the next candidate from the same state. The downloadable records include the directory listing, listed city and state, inspected URL and individual observations.
State assignment follows the listed office address, which may differ from a brokerage’s headquarters. One directory-linked page was inspected per website; some were location or contact pages. The sample includes local agencies and larger brokerages, and separate websites may share corporate ownership.
Equal state counts provide geographic balance. They do not weight results by population or number of agencies. Directory membership and accessible page source also shape the sample. The audit measures source elements, not loading speed, completed calls, form delivery or sales, and four websites per state are insufficient for state rankings.
The methodology explains the selection and measurement rules. The selection log records considered candidates and exclusions so readers can inspect how the sample was assembled.