The audit as it stood in June 2026

Your robots file blocks CCBot, which is one of the sources the AI tools train and retrieve on.

Remove the CCBot disallow rule unless the block is deliberate. It is more often left over from a template than chosen.

3 important pages are tagged "hide this page from Google" (noindex), so they will not appear in Google.

Ask your web developer to remove the noindex tag from pages you want found. Two of these are practice-area pages that should be among your strongest.

Your firewall blocks the crawlers ChatGPT and Perplexity use, so neither can read any page on the site.

Ask whoever manages your hosting or CDN to allow GPTBot, ClaudeBot, PerplexityBot and CCBot. Your robots file already permits them; the block is happening before that file is ever read.

No robots.txt rule blocks Google from the site or its CSS and JavaScript files.

No page is accidentally tagged noindex, so nothing important is hidden from Google.

No X-Robots-Tag server header quietly tells Google to skip pages.

No page failed with a 5xx server error.

Every checked page loaded over HTTPS with the secure padlock connection.

The SSL certificate was valid and correctly installed for the site.

The robots.txt file is free of formatting mistakes.

The site has an XML sitemap and it loads correctly.

The XML sitemap follows the format Google requires.

Every URL listed in the sitemap actually loads.

The sitemap lists only indexable pages meant to appear in Google.

Every page has a canonical tag naming its official version.

Each normal page's canonical tag points to itself.

No pages send Google conflicting canonical signals.

No pages returned a 404 not found error for visitors across the 200 pages crawled.

Any moved pages redirected in one hop with no redirect chains found in the crawl.

No redirect loops were found, so visitors were not sent around in circles.

Permanently moved pages used 301 redirects, so ranking credit could transfer properly.

Every internal link led to a working page across the site.

No orphan pages were found, so every page had at least one internal link pointing to it.

Every page was reachable within three or four clicks of the homepage.

The SSL certificate had more than 30 days left before expiring.

Insecure http:// addresses redirected to the secure https:// version of the site.

No mixed content was found, so secure pages loaded nothing over insecure connections.

The site used one preferred version of the address, and the other version redirected to it.

The paginated pages were set up so Google could read the numbered page sequence correctly.

No two pages were found with word-for-word identical content across the 200 pages crawled.

The location pages were not just the same template with the city name swapped.

No two pages shared the same title tag across the pages checked.

No two pages shared the same meta description across the pages checked.

The site did not use hreflang tags on a single-language site.

The key text and links were present in the raw HTML, not added later by JavaScript.

Missing pages returned a real 404 instead of a soft 404.

robots.txt, meta robots tags, and the XML sitemap all agreed on which pages should be indexed.

The robots.txt file exists and loads on the site.

The robots.txt file points Google to the XML sitemap.

The sitemap lastmod dates look honest and not auto-generated.

The sitemap stays within Google's 50,000-URL and 50 MB limits.

Canonical tags use complete, secure https addresses written consistently.

Every page's HTTP status was recorded as fine, moved, or broken.

All outbound links to other websites were working when checked.

All checked URLs stayed reasonably short and did not run unusually long.

The site’s URLs were free of tracking parameters and session IDs.

The site’s URLs used hyphens to separate words, not underscores.

The site’s URLs stuck to lowercase letters.

The site’s URLs avoided spaces and special characters.

The site’s URLs were consistent about the trailing slash.

The HSTS header was in place, telling browsers to always use HTTPS.

Filter and sort URLs did not create a flood of near-duplicate pages.

No two pages shared the same H1 headline across the pages checked.

Any hreflang language tags in use were set up correctly.

Each page's html lang attribute matched the language of its content.

No render-blocking CSS or JavaScript was found holding up the pages from appearing.

A helpful custom 404 page greeted visitors who followed a dead link.

URL parameters did not multiply into countless duplicate pages.

Blocking GPTBot, ClaudeBot, and CCBot appeared to be a deliberate choice rather than an accident.

This does not apply to your firm.

This does not apply to your firm.

41 images across 12 pages have no alt text, so screen readers and search engines see nothing where those images are.

Ask whoever edits the site to add a short description to every meaningful image.

19 pages have no Twitter Card tags.

Minor, but it means shared links show as bare URLs instead of a headline and picture.

Every one of the 200 crawled pages has a title tag for its search result headline.

Each of the 200 crawled pages has a unique title tag.

The target keyword appears early in the title tag on each of the 200 crawled pages.

Every one of the 200 crawled pages has a meta description for its search result blurb.

Every one of the 200 crawled pages has a non-empty H1 main headline.

The H1 headline on each of the 200 crawled pages names the page topic clearly.

No content page in the 200-page crawl was thin, under roughly 300 words.

Each page in the 200-page crawl had enough depth for its job.

Each page’s content matched what its title promised across the 200 pages checked.

The site keeps image files small enough to avoid slowing down the 200 pages checked.

The practice-area and contact pages received internal links from around the site.

Content pages showed a visible last-updated date.

The main service pages answered common client questions right on the page.

The title tags on the 200 crawled pages fit Google’s 30 to 60 character display range.

The firm name appears in the title tags across the 200 crawled pages.

The meta descriptions on the 200 crawled pages fit Google’s 120 to 160 character display range.

Each of the 200 crawled pages has a unique meta description written for that page.

The meta descriptions across the 200 crawled pages give searchers a reason to act.

Each of the 200 crawled pages has exactly one H1 main headline.

The headings on the 200 crawled pages follow a logical outline without skipped levels.

The 200 crawled pages read naturally without keyword stuffing.

The alt text on the 200 crawled pages genuinely describes the images.

Images on the 200 pages checked declared their width and height, so the layout stayed stable while loading.

The site used modern image formats like WebP or AVIF on the 200 pages checked.

Images below the fold were set to lazy-load on the pages checked, so they waited until needed.

Image files used descriptive filenames on the pages checked instead of camera-style codes.

The 200 crawled pages had a healthy internal-link count, with pages linking to other pages on the site in a balanced way.

Link text on the 200 pages checked said where the link led, instead of using vague phrases like click here.

The content linked out to credible outside sources where it helped the reader on the pages checked.

The page URLs used short, descriptive slugs on the 200 pages checked.

The pages checked had Open Graph tags for social sharing previews.

The og:image share picture loaded at the right size on the pages checked.

The site had a favicon in place for browser tabs and mobile search results.

The writing was easy to read for the firm's audience on the pages checked.

Long articles opened with a table of contents that matched the sections on the page.

Key information sat in plain view rather than being hidden behind tabs or accordions on the pages checked.

The site describes itself with the deprecated "Attorney" schema type.

Switch to LegalService. Attorney was retired and is no longer understood reliably, which costs you the machine-readable description of what the firm does.

The video pages carry no VideoObject markup.

Adding it lets the videos appear as video results rather than plain links.

The pages carried structured data in JSON-LD, the format Google prefers.

The site's JSON-LD structured data parsed without syntax errors across the pages checked.

The homepage carried Organization schema identifying the firm.

The sameAs fields linked the firm's official profiles to one identity in the structured data checked.

Each attorney bio carried Person schema with title and bar membership details.

Each of the 6 practice-area pages had its own Service schema.

The site's structured data validated with zero structural errors.

The name, address, and phone in schema matched the visible site details.

Business hours were declared in schema and matched the firm's actual hours.

The schema claims matched what visitors can actually see on the pages checked.

The pages with genuine question-and-answer sections carried FAQPage schema that AI tools can read.

The breadcrumb schema matched the visible navigation trail on the pages checked.

Review and star-rating schema was used only for reviews shown on the firm's own site.

The article pages carried Article schema with author and date details.

The site's WebSite schema declared the firm's official name to Google.

Recommended schema details were mostly filled in, with only a few warnings left.

Each schema type included both its required and recommended fields.

The LocalBusiness schema included geo coordinates for the office.

The schema spelled out the areas the practice serves.

The site’s schema cross-references connect the firm and its attorneys as one organization across the pages checked.

The site’s markup did not rely on retired schema types for enhanced listings.

This does not apply to your firm.

No verified Google Business Profile was found for the Katy office.

Claim and verify a profile for the Katy address. Without one the office cannot appear in the map results, and the AI tools have nothing to read about that location.

The Houston profile’s primary category is "Law firm", the most generic option available.

Change the primary category to "Divorce lawyer" or "Family law attorney". Primary category is the strongest single lever in local ranking, and a generic one competes with every practice area in the city.

The street address on the Google Business Profile doesn't match what the website shows. Mismatched listings confuse Google about which details are real and cost map ranking.

Make the name, address, and phone identical on the website and the Google Business Profile; pick one canonical format and use it everywhere.

The Katy office has no Google reviews.

Reviews are both a ranking factor and the thing a prospective client scans first. A location with none reads as a location that is not really open.

The Houston profile has no description and only 4 photos.

Fill in the description and add photos of the office and the team. Both are read by prospective clients and by the tools.

Only one secondary category is set on the Houston profile.

Add the other services you genuinely offer. Each accurate category is another local search you can appear in.

The Katy page is 88 percent identical to the Houston page.

Write the Katy page for Katy: the courthouse, the local process, the questions clients there actually ask.

You serve 2 cities across 2 practice areas but have pages for only 1 of the 4 combinations.

Estate planning in Katy has no page at all, which is why the AI tools never name you for it.

Your phone number differs across 3 directory listings.

Pick one number and make every listing use it. Two of these still show a number that was retired.

Reviews arrive in bursts rather than steadily: across the 14 months covered by the 87 most recent reviews we sampled, only 4 months saw any review (3 in the last 90 days). Bursts read as campaigns, to Google and to clients.

Make the review ask part of closing every matter so reviews arrive steadily instead of in bursts.

You have replied to 12 of the 87 reviews sampled.

Replies are read by everyone who comes after. A short, courteous reply to a poor review often matters more than the review.

Reviews live only on Google and are not shown anywhere on the site.

Someone who reaches your site should not have to leave it to find out that other people recommend you.

The site’s name, address, and phone read the same everywhere checked.

The site had a dedicated location page for each office in Houston and Katy.

The firm was listed in the major legal directories.

The site’s titles and headlines targeted service and city phrases like estate planning attorney Houston.

The site covered its services and locations well enough to compete for near me searches.

The Houston Google Business Profile had a 4.6 star average rating.

The site spelled out its service area in page text and schema.

The contact and location pages included an embedded Google Map.

The structured data included map coordinates for the office locations.

Driving directions and a one-tap map link were easy to find.

The site gave visitors a location finder for its two offices.

This does not apply to your firm.

Largest Contentful Paint (how quickly the main content appears) is poor for real visitors: 4.61s (target under 2.5s).

Ask your web developer to make the main content load faster: smaller top-of-page images, quicker server responses.

Mobile speed has drifted from 71 to 58 over the last six months.

Nothing here is broken yet, but the trend is one direction. Most of it is a chat widget and three tracking scripts.

Interaction to Next Paint (how fast the page responds when tapped or clicked) is good for real visitors: 142ms (target under 200ms). By device: mobile 168ms, desktop 96ms.

Cumulative Layout Shift (how much the page jumps around while loading) is good for real visitors: 0.04 (target under 0.1). By device: mobile 0.05, desktop 0.02.

All three Core Web Vitals passed together.

Time to first byte showed the server started responding within a healthy window.

The site’s pages stayed light enough for phone connections across the 200 crawled pages.

The images were compressed, modern-format, and sized to fit their space across the 200 crawled pages.

Text compression was enabled, so pages were shrunk before being sent on the 200 crawled pages.

Chat widgets and other third-party scripts did not drag down speed across the 200 crawled pages.

First Contentful Paint was measured and showed the page began rendering promptly.

Total Blocking Time was measured and stayed within a healthy range.

Speed Index was measured in lab testing and showed the page filled in quickly.

The Lighthouse performance score was 58 on mobile and 81 on desktop.

Speed was checked in both Lighthouse lab tests and CrUX real visitor data.

Speed was measured separately for mobile and desktop, with mobile scoring 58 and desktop 81.

CrUX real visitor speed data was collected page by page where enough traffic existed.

The pages did not fetch an excessive number of HTTP requests across the 200 crawled pages.

No render-blocking files were found across the 200 crawled pages.

The pages did not ship unused CSS across the 200 crawled pages.

The pages did not ship unused JavaScript across the 200 crawled pages.

Browser caching was set up so returning visitors could reuse saved files on the 200 crawled pages.

A CDN served the site from nearby data centers, with Cloudflare in front of the WordPress site.

Font loading was optimized so text showed immediately across the 200 crawled pages.

The site was responsive and rearranged itself properly for phone and tablet screens.

The viewport tag was present and set correctly for phones.

Phone visitors saw the same content desktop visitors did across the 200 crawled pages.

The pages passed Core Web Vitals on mobile.

No intrusive pop-ups smothered the content on a phone screen.

The phone number was a tap-to-dial link on phones.

Buttons and links were big enough and spaced well enough to tap comfortably on phones.

Text on phones was large enough to read without zooming.

A person on a phone could reach the firm in a few easy steps.

No leftover AMP pages were found across the 200 crawled pages on the WordPress site behind Cloudflare.

No "Attorney Advertising" label appears on the site.

Add the label your state bar requires, usually in the footer. This is a compliance item rather than a marketing one.

Case results are shown on 4 pages with no disclaimer that past results do not guarantee a future outcome.

Add the disclaimer wherever outcomes are described. Most state bars require it, and it is the kind of omission that is cheap to fix and expensive to be caught on.

11 practice-area pages open with background rather than an answer, so there is no passage an AI tool can lift.

Start each page with two or three sentences that answer the question in its title, then go into detail. This is the single change that most affects whether you get quoted.

23 blog posts carry no author and no reviewer.

Legal content without a named attorney behind it is weighted down by Google and skipped by the AI tools. Attribute them to whoever wrote or approved them.

Two pages describe the firm as "specialists" without naming a board certification.

Either name the certification or change the wording. Several state bars restrict the word.

The site stated that reading it or sending a message does not create an attorney-client relationship.

Every article had a named author byline across the 41 blog posts checked.

Professional credentials were shown near the content for all 3 attorneys.

The content cited real authorities such as statutes, cases, courts, or bar sources where needed.

The About page gave a full picture of the firm, including the team, credentials, history, and Houston and Katy offices.

The Contact page included the street address, phone, email or form, and a map for both offices.

The site had a privacy policy.

The firm's physical office address was shown on the site.

The core service pages treated their topics in real depth across the 6 practice-area pages.

Related content was organized into interlinked topic clusters around the main practice-area pages and blog posts.

Content showed visible last-reviewed dates to indicate it is actively maintained.

The content appeared original and not copied or lightly reworded from other sites.

The pages were written to help people first, not just to rank in search results.

The site’s key content was server-rendered in the HTML, where AI tools could read it.

Each author had a bio page describing their background and qualifications for all 3 attorneys.

Author Person schema named the attorneys and their credentials across the site.

The site had a terms of service page.

The site’s AI-assisted content was expert-reviewed, original, and signed by a named author.

Each page matched the search intent of the people it was meant to reach.

Two competitors have more than three times as many websites linking to them (Carmody Legal Group 307, Reyes & Whitfield 323, against your 96).

This gap takes months rather than weeks to close, and it is the main reason those two outrank you on the searches you share.

The average spam score of the sites linking to you is fine (14/100), but 2 of the 100 strongest linking domains carry a high spam score, led by legal-seo-links.example (78/100) and lawdirectorynow.example (62/100).

Ask your marketing team to review these sites and disavow the ones that are junk (tell Google to ignore those links).

The site had 96 referring domains linking to it.

The anchor text from other sites looked natural.

The site had backlinks from respected institutions in the field.

The site had 412 backlinks pointing to it.

The site had a natural mix of dofollow and nofollow links.

The site’s domain authority was tracked as a benchmark for overall strength.

Your links are stable or growing: 38 new linking sites appeared while 21 dropped off (across the past year of monthly snapshots in our link index).

We could not measure this on the last run.

We could not measure this on the last run.

Google shows a map pack (local results box) for the sampled search "divorce lawyer houston", and Alvarado Family Law is not in it.

Map-pack placement is driven by your Google Business Profile: the right primary category, a complete listing, reviews, and consistent name/address/phone.

Google’s AI Overview appears for your main search and cites Reyes & Whitfield and two directories, not you.

The pages it cites all answer the question directly in their opening lines. Yours do not, which is the same problem as the answer-first finding above.

1 search term has two or more of your pages ranking for near-identical variants of it ("houston divorce lawyer"), splitting the ranking between them. Judged from real ranking data.

Ask whoever writes your content to combine the competing pages, or differentiate what each targets, so they stop splitting one term's ranking.

14 keywords sit at positions 11 to 20, just off the first page.

These are the fastest wins on the site. Strengthening the pages that already rank here is cheaper than starting new ones.

The site’s Google rankings for the main keywords were identified.

The rankings were spread across page one, page two, and beyond.

Monthly search volume was listed for each target keyword.

Keyword difficulty was estimated for each ranked keyword, shown per keyword in the rankings table above.

The audit listed the search phrases the site appears for, including both targeted and unexpected terms.

The audit split search traffic into branded and non-branded visits, showing both parts separately.

The site was compared head-to-head with named competitors for the searches they both target.

We could not measure this on the last run.

Google Safe Browsing did not flag the site for malware or phishing.

The TLS encryption behind the site’s padlock used modern, strong settings.

The domain’s DNS records were sound for routing the website and email.

The server answered promptly when pages were requested.

The site stayed up and reachable throughout the audit.

The audit captured the domain’s age and registration history.

Images carried alt text across the 200 pages checked.

Every link had text a screen reader could announce.

Text contrast met the WCAG accessibility standard.

Every form field had a label screen readers could announce.

The site’s Lighthouse accessibility score was captured.

Google Analytics 4 was installed to measure site traffic.

The site has Google Tag Manager in place for measurement tags.

The site is connected to Google Search Console, Google’s free search reporting dashboard.

A cookie consent notice appears where privacy rules require one.

The site includes working links to the practice’s social profiles.

The site’s Organization schema lists the practice’s social profiles in sameAs.