1 in 3 Sacramento Small-Business Homepages We Tested Had No Tap-to-Call Link

Out of 45 fully measured homepages unless noted:

• No tap-to-call link: 15 of 45 (33%).
• Phone number shown but not tappable: 10 of 45 (22%), about 1 in 5.
• No contact form and no booking, reservation, ordering, quote or estimate link: 12 of 45 (27%), more than 1 in 4.
• No form, no booking or quote link and no tap-to-call link at all: 7 of 45 (16%).
• Slow main content: 38 of 42 scored homepages (90%) took more than 4 seconds to show their main content. The median was 9.2 seconds.
• No "good" speed scores: 0 of 42 reached 90. The median Lighthouse mobile score was about 53 out of 100.
• Heavy pages: 16 of 45 (36%) weighed more than 3 MB. The median was about 2.1 MB.
• No business details for Google: 18 of 45 (40%) had no LocalBusiness or Organization schema.
• Old copyright year: 8 of the 34 homepages that show one (24%) showed 2024 or earlier.
• Secure connection: 51 of 51 reachable sites with a working domain (100%) loaded over HTTPS.

How we ran this study

• Sample: 60 businesses drawn at random (with a fixed seed) from 101 eligible independent local businesses in a public Sacramento business directory. Before the draw we screened out chains, franchises, national or multi-state firms, large regional firms, nonprofits, public agencies, marketing agencies and our own clients.
• After the screen: 1 business was removed after the draw because it failed the local small business screen. Sample: 59.
• Not measured (14 of 59): 2 listings link to a web address that does not exist. 5 sites closed the connection to our test machine during the secure handshake (all 5 loaded normally through a separate web fetch service, so we count this as a limit of our network, not a broken site). 6 sites showed our headless browser a hosting company's "robot challenge" page instead of the homepage on 2 or 3 tries. 1 listing forwards to another sampled business's website (see the footnote).
• Fully measured: 45 homepages. Lighthouse-scored: 42 of those 45.
• When: Wednesday, October 7, 2026, about 1:10 to 2:25 AM PT.
• Tools: curl with an iPhone Safari user agent and certificate checking on; headless Google Chrome at phone size (390 x 844); Lighthouse 12 mobile (simulated mid-range phone on slow 4G), run twice per site, using the median of the 2 runs.
• Mix of the 59: food and drink 14, security services 8, construction and trades 7, legal 6, health and wellness 6, retail 5, printing and media 5, accounting 4, real estate 3, beauty and personal care 1.
• Limits: homepage only, one point in time, lab speed data and a small sample. See the limits section below.

Footnote: one sampled listing's web address forwards to a page on the website of another business in the sample. We kept it in the 59 and in the directory link count, but left it out of every page measurement and the HTTPS count, so one website was not measured twice.

Most people who look up a local business on a phone want to call, book or ask for a price. We checked whether the homepage lets them.

• A tap-to-call link (a real phone link you can tap) was on 30 of 45 homepages (67%). That leaves 15 of 45 (33%) with none.
• A phone number was visible as text on 37 of 45 (82%).
• On 10 of 45 (22%), the number was there but not tappable. The customer has to memorize it or copy it by hand.
• A contact form on the homepage itself: 12 of 45 (27%).
• A booking, scheduling, reservation, ordering, quote or estimate link, or an embedded booking widget: 24 of 45 (53%).
• Any way to act from the homepage (a form or one of those links): 33 of 45 (73%). The other 12 of 45 (27%) had neither.
• 7 of 45 (16%) had no form, no booking or quote link and no tap-to-call link.

Our check is strict. A button that only says "Get started" or "Contact us", or a plain link to a contact page, did not count, so "any way to act" is a floor. The tap-to-call number needs no interpretation: the phone number is a link or it is not.

We used Lighthouse 12 in mobile mode, which simulates a mid-range phone on slow 4G. Each site ran twice and we used the median of the 2 runs, which for two runs is their average. 40 sites had 2 valid runs and 2 had 1 valid run. 3 measured homepages could not be scored because Lighthouse was shown the hosting company's challenge page.

• Median performance score: about 53 out of 100 (52.75).
• Poor (under 50): 19 of 42 (45%). Needs improvement (50 to 89): 23. Good (90 or more): none on the two-run median. 3 sites hit 90 or more on one run but not the other.
• Largest Contentful Paint (the moment the main content shows): median 9.2 seconds. Only 1 of 42 was at or under Google's 2.5 second "good" mark. 38 of 42 (90%) took more than 4 seconds, which Google calls poor, and 20 of 42 (48%) took more than 10 seconds.

Lab scores bounce around, which is why we ran every site twice. The two runs of the same site differed by a median of 3.5 points, and 3 of 40 differed by more than 20 points. We could not use Google PageSpeed Insights because its free API returned a daily quota error, so these scores come from Lighthouse on our own machine and will not match PageSpeed Insights exactly.

One thing that was not the problem: the servers. The median time to first byte was 0.24 seconds from our machine, and only 2 of 45 took over 1 second. The slowness is mostly page weight and scripts, not slow hosting.

Page weight is how much data a phone has to download to show the homepage. We measured it in headless Chrome at phone size.

• Median homepage size: about 2.1 MB (2,180 KB). The smallest was 220 KB. The largest was about 54 MB.
• Over 3 MB: 16 of 45 (36%). Over 5 MB: 8 of 45 (18%).
• Network requests: the median homepage needed 70 requests to load, and 17 of 45 (38%) needed more than 100.

Each request is a photo, font, script, widget or tracking tag the phone has to fetch. A heavy page can feel fine on office Wi-Fi and slow on a phone in a parking lot.

This is where Sacramento small businesses did well.

• Secure connection: 51 of 51 reachable sites with a working domain (100%) loaded over HTTPS with a valid certificate.
• Forwarding to the secure version: plain http:// addresses forwarded to https:// on 41 of 42 we could check (98%).
• Mobile viewport tag: present on 41 of 45 (91%). The other 4 of 45 (9%) lack it, which usually means a shrunken desktop page on a phone.

Google reads the page title and any structured data (schema) that spells out who the business is.

• Page title: present on 45 of 45 (100%). The median title was 49 characters. 8 of 45 were over 60 characters, so they are likely cut off in Google results, and 11 of 45 were under 30, which leaves little room for a service and a town.
• LocalBusiness or Organization schema: on 27 of 45 (60%). 18 of 45 (40%), about 4 in 10, gave Google no structured business details on the homepage.

Our schema check only looks for the business types. It does not judge whether the details inside are complete or correct, so 60% is the share with schema, not the share with complete schema.

A footer copyright year is a small upkeep signal. An old year does not mean a business is closed or a site is broken, but it is an easy line to keep current.

• A copyright year was shown in the page text on 34 of 45 homepages.
• Of those 34, 8 (24%) showed 2024 or earlier. The oldest was 2018.
• 24 of 34 showed 2026.

We also checked the directory listings. 2 of 59 (3%) link to a web address that does not exist; one looks like a typo. Every click on a dead address is lost.

Here is the order we would work in.

1. Make your phone number a tap-to-call link

The cheapest fix on this list. Open your homepage on your phone and tap the number. If nothing happens, ask your web person to make it a phone link in the header and footer. If the number is inside an image, add it as real text too.

2. Give every visitor one clear next step

More than 1 in 4 homepages had no form and no booking or quote link. Pick the action you want most (call, book or request a quote) and put it near the top, with a button that says what happens. If you send ad or social traffic to your homepage, read our post on landing page vs homepage for when a page built for one offer works better.

3. Put your homepage on a diet

Speed is the second big leak. Start with photos (resize them, use modern formats), then autoplay video at the top, then widgets and tracking tags you no longer use. Our 5-minute speed test walks through how to check your own site and read the numbers. Run it twice, because scores move.

4. Fix the quick search basics

Check that your page title says what you do and where, in about 60 characters or less. Ask whether your homepage has LocalBusiness or Organization schema, and that the name, address and phone in it match your Google profile. Our Google Business Profile checklist covers keeping those details consistent.

5. Do the five-minute upkeep pass

Update the footer year. Then check every place that links to your site (directories, your Google profile, social bios) and make sure each one lands on a working page.

When it is more than a fix

If your site has no viewport tag, is heavy on every page or breaks when you edit it, patching it item by item can cost more than starting over. That is the kind of decision we help with for website design in Sacramento. If you want a second set of eyes first, our free audit looks at your site on a phone and tells you what to fix first.

• One source. Every business came from one public Sacramento business directory, and listed businesses may be more established than average. The findings describe this sample, not every business in the region.
• Small sample. One site moves a percentage by about 2 points. Read "33%" as "about 1 in 3".
• 14 of 59 not measured. If those differ from the rest, the results could shift. A challenge page is a sign of strict hosting security, not of a bad site.
• Judgment calls. The small business screen was applied by hand, using written rules.
• Homepage only. A site may have booking, schema or a phone link deeper in.
• Pattern checks. Phone numbers inside images are missed. We hand-checked 6 randomly picked sites against their raw page code for tap-to-call, schema and copyright year, and 18 of 18 checks agreed.
• Lab speed, one point in time. Lighthouse ran from one machine on a data center connection, and sites change.

• Vessel Archive study data: 59 sampled businesses, 45 homepages fully measured and 42 Lighthouse-scored on October 7, 2026. Totals only; per-site data is not published.
• PageSpeed Insights, About PageSpeed Insights (90 or above good, 50 to 89 needs improvement, below 50 poor)
• Chrome for Developers, How Lighthouse calculates your overall Performance score
• web.dev, Largest Contentful Paint (good 2.5 seconds or less, poor over 4.0 seconds)
• Google Search Central, Local business (LocalBusiness) structured data