Original research · 117 Mesa websites · tested October 31, 2026
We tested 117 Mesa small business websites the way a customer's phone sees them.
Every site was loaded on a simulated phone with Google Lighthouse and then inspected for the things that turn a visit into a call: a tap-to-call number, a form, a way to book. The median site scored 61 out of 100 for mobile performance. The rest of the story is below, in aggregate, with the method attached.
Key findings
117 Mesa websites, measured. Medians and shares, not anecdotes.
What we tested
117 independent Mesa businesses. One homepage each.
We started with 170 candidate websites found through ordinary web searches for Mesa businesses by trade, then kept only independent businesses that publish a Mesa, Arizona street address on their own site. No franchises, no national chains, no lead-generation sites, and nothing we built ourselves. Each homepage was tested on October 31, 2026.
We publish aggregates only. No business is named, no single site’s score is shown, and no screenshot of anyone’s site appears here. The point is the pattern, not the scoreboard.
- Home services22% (26)
- Health22% (26)
- Professional services16% (19)
- Food & drink10% (12)
- Auto repair9% (10)
- Salons & barbers10% (12)
- Retail10% (12)
Finding 1 · Speed on a phone
The median Mesa homepage scored 61 for mobile performance.
14% of the 104 sites Lighthouse could test landed in the poor band (below 50). Only 13% reached 90 or better. The middle half of the sample scored between 55 and 72.
The slow part is the first impression. The median site needed 10 seconds to paint its largest element on Lighthouse’s simulated mid-range phone and slow connection. That is a stress test, not a stopwatch on your customer’s phone; on fast Wi-Fi the same page will usually arrive sooner. But every site faced the same test, and only 6% met the 2.5-second mark Google uses for a good Largest Contentful Paint. Once loaded, most pages behaved: 90% kept layout shift within Google’s target and 70% kept script blocking at 200 ms or less.
Mobile performance scores
Number of sites in each 10-point band
Lighthouse performance score
Time until the main content appears
Largest Contentful Paint, simulated phone on a throttled connection
- ≤2.5 s6% (6)
- 2.5–4 s9% (9)
- 4–6 s14% (14)
- 6–8 s14% (14)
- 8–10 s9% (9)
- 10–15 s20% (21)
- >15 s30% (31)
6% of sites came in at 2.5 s or less.
Time the page was frozen by scripts. 5% of sites exceeded 600 ms.
25% of homepages weighed more than 4 MB.
LCP 2.5 s or less, layout shift 0.1 or less and blocking time 200 ms or less, in the lab.
Page weight
- Under 1 MB12% (12)
- 1–2 MB26% (27)
- 2–4 MB38% (39)
- 4–8 MB17% (18)
- Over 8 MB8% (8)
Layout shift
- 0.1 or less90% (94)
- 0.1–0.257% (7)
- Over 0.253% (3)
Finding 2 · Can a customer reach you?
Most sites let you tap to call. Few let you book.
83% of homepages had a working tap-to-call link. 31% had a lead form on the homepage itself, either built into the page or embedded from a form service. Just 17% offered a way to book online, and 10% had neither a tap-to-call link nor a homepage form. Only 6% offered all three.
- Tap-to-call (tel:) link83% (97 of 117)
- Lead form on the homepage31% (36 of 117)
- Chat widget9% (10 of 117)
- Online booking (widget or link)17% (20 of 117)
- Call, form and booking, all three6% (7 of 117)
By industry
| Group | Sites | Tap to call | Form | Booking |
|---|---|---|---|---|
| Home services | 26 | 96% | 46% | 8% |
| Health | 26 | 89% | 31% | 39% |
| Professional services | 19 | 90% | 32% | 0% |
| Food & drink | 12 | 83% | 8% | 8% |
| Auto repair | 10 | 100% | 30% | 20% |
| Salons & barbers | 12 | 50% | 8% | 42% |
| Retail | 12 | 50% | 42% | 0% |
Groups are small; treat differences of a site or two as noise. Booking services detected: Boulevard (2), Zenoti (2), Tekmetric (2), Vagaro (2), Calendly (1), Jobber (1), NexHealth (1), ChiroTouch (1), Jane (1), Mangomint (1), Meevo (1), Aesthetic Record (1), Toast Tables (1), Phorest (1), Booksy (1), Square Appointments (1).
Finding 3 · Can Google read you?
The search basics are mostly there. The details that help local search often are not.
Every site had a page title and a mobile viewport tag, 98% loaded over HTTPS, and the median Lighthouse SEO score was 92. But only 58% described the business in structured data as a LocalBusiness (or one of its subtypes, such as Dentist or Plumber), the machine-readable format search engines use to connect a site to a name, address and hours. 14% had no h1 heading at all, and 27% had more than one.
- Served over HTTPS98% (115 of 117)
- Mobile viewport tag100% (117 of 117)
- Page title100% (117 of 117)
- Meta description94% (110 of 117)
- Any JSON-LD structured data81% (95 of 117)
- LocalBusiness structured data58% (68 of 117)
- Exactly one h1 heading60% (70 of 117)
Lighthouse SEO scores
- Poor (0–49)0% (0)
- Needs work (50–89)24% (25)
- Good (90–100)76% (79)
29% of page titles ran longer than 60 characters, a common rule of thumb for where search results start cutting them off. 94% had a meta description, which leaves 6% letting the search engine write its own snippet.
Finding 4 · Can everyone use it?
63% of sites had text with too little contrast to read comfortably.
The median accessibility score was 90, and 51% of sites scored 90 or better. Automated checks catch only part of what makes a site usable with a screen reader or a shaky thumb, so treat these as a floor, not a verdict. Among the 41 sites with a visible form field on the homepage, 46% had at least one field with no label a screen reader could announce (63 of 209 fields overall).
Most common accessibility failures
Share of sites where the check applied and failed
- Text with too little color contrast63% (65 of 104)
- Links with no discernible name54% (56 of 104)
- Headings that skip levels50% (51 of 103)
- Images missing alt text19% (20 of 104)
- Tap targets too small or too close19% (20 of 104)
- Form fields without labels (Lighthouse check)16% (6 of 37)
- Buttons with no accessible name11% (8 of 71)
- Page missing a language attribute4% (4 of 104)
Accessibility scores
- Poor (0–49)0% (0)
- Needs work (50–89)49% (51)
- Good (90–100)51% (53)
Best-practices scores
- Poor (0–49)0% (0)
- Needs work (50–89)58% (60)
- Good (90–100)42% (44)
Finding 5 · What the sites are built on
WordPress runs 43% of the sample.
Platform was identified from the page code (WordPress theme paths, builder asset hosts, generator tags). “Other or custom” covers hand-built sites, industry platforms and anything we could not identify with confidence.
- WordPress43% (50)
- Other or custom27% (32)
- Squarespace7% (8)
- Duda7% (8)
- Wix4% (5)
- GoDaddy Website Builder4% (5)
- Webflow3% (4)
- Shopify3% (4)
- Weebly1% (1)
Median mobile performance by platform
| Platform | Sites tested | Median score | Median LCP |
|---|---|---|---|
| WordPress | 46 | 61 | 10 s |
| Other or custom | 31 | 62 | 9.7 s |
| Squarespace | 8 | 49 | 16.6 s |
| Wix | 5 | 75 | 6.1 s |
| GoDaddy Website Builder | 5 | 69 | 6.1 s |
Platforms with fewer than five Lighthouse results are left out. All 8 sites on the Duda builder refused Lighthouse's request, so Duda has no speed figures here. With groups this small, read the differences as hints, not rankings.
What it means for a business owner
Five fixes, roughly in order of cost.
None of this requires a new website. Most of it is an afternoon. Here is what the data says to check first, cheapest first.
01Make the phone number a button.
Wrap every phone number in a tel: link, and put one in the header where a thumb can find it. 17% of the sites we tested had not. It is a five-minute change on almost any platform.
02Test your homepage on a phone, not your desk.
Run the free Website Health Check or PageSpeed Insights in mobile mode. If your score sits below this report’s median of 61, start with images: compress them and serve them at the size they display. Our PageSpeed guide for trades covers the usual culprits, and a flat-priced Speed Tune handles the rest.
03Label every form field.
A placeholder that vanishes when you type is not a label. Give each field a real one, so screen readers and autofill both know what goes where. If the form has stopped sending altogether, that is a different and more urgent problem.
04Tell search engines who and where you are.
Add LocalBusiness structured data with your name, address, phone and hours, matching your Google Business Profile exactly. 42% of the sample had not. Our Map Pack guide explains why consistency matters, and local SEO is where we do it for clients.
05Let people book without calling.
Only 17% of sites offered online booking. In our experience, the customer who finds you at 9 p.m. is likelier to pick a slot than to leave a voicemail. Most scheduling tools embed in minutes; every website package we build includes online booking, and an AI receptionist can answer the calls nobody is there for.
If the list reads like a description of your whole site, a rebuild may cost less than the patching. Our redesign service and conversion work start from exactly these checks, and The Leads You Never Knew You Lost explains what a missing call button costs in practice.
Methodology
How the numbers were made.
Sample selection
We ran public web searches for Mesa, Arizona businesses in each target trade (for example “Mesa AZ plumber” or “dentist Mesa AZ”) and recorded the business websites in the order they appeared, skipping directories, review sites, marketplaces and listings for known franchises and national chains. That produced 170 candidates. A site stayed in the sample only if it published a Mesa, Arizona address: in the homepage text with a Mesa ZIP code (99 sites), in the homepage’s structured data (12), or on a contact or about page linked from the homepage (6). Of the 170 candidates, we excluded:
- No Mesa, AZ street address on the site: 24
- Lead-generation or referral site, not a business: 3
- Bot check blocked automated measurement: 5
- Offline: domain did not resolve: 3
- Expired, parked or blank: 2
- Over the per-category cap of 12: 16
To keep a single category from dominating, each category was capped at twelve sites, keeping those found first. Independent businesses with a handful of Phoenix-area locations were kept; real estate agents and teams that work under a larger brokerage were kept as independent businesses. The final sample is 117 sites; Lighthouse completed a valid run on 104 of them.
Dates and tools
- Test window: October 31, 2026 (first test 2026-10-31 08:55 UTC, last 2026-10-31 10:23 UTC).
- Google Lighthouse 12.6.0, run from the command line: mobile form factor (412 x 823 screen emulation), Lighthouse default simulated throttling (150 ms round trip, 1.6 Mbps throughput, 4x CPU slowdown), headless Chrome, one run per site. Categories: performance, accessibility, best practices, SEO.
- Google Chrome 155.0.8059.41, headless.
- Page checks: Headless Chrome via puppeteer-core 24.10.0, mobile viewport 412x823, page rendered and inspected after load. Node.js v24.18.0.
- Lighthouse runs were strictly sequential, one site at a time with a pause between sites. Page checks ran in up to three browser sessions at once, never more than one per site.
- Scores follow Google’s published Lighthouse performance scoring.
Definitions
- Performance, accessibility, best-practices and SEO scores: Lighthouse category scores, 0–100. Bands follow Google’s: 0–49 poor, 50–89 needs improvement, 90–100 good.
- LCP, CLS, TBT: Largest Contentful Paint, Cumulative Layout Shift and Total Blocking Time from the same Lighthouse run. These are lab values under simulated throttling, not field data from real visitors.
- Page weight: Lighthouse’s total byte weight for the homepage load.
- HTTPS: the homepage finished loading on an https:// address.
- Mobile viewport tag: a viewport meta tag that sets a width or an initial scale.
- Tap-to-call: at least one link whose address begins with tel: on the rendered homepage.
- Lead form: a form containing an email field, a phone field or a message box, or an embedded third-party form (such as a form-builder iframe). Search boxes and newsletter-only fields without email do not count.
- Unlabeled field: a visible input, select or text area with no associated label element, aria-label, aria-labelledby or title. Placeholder text does not count as a label.
- Online booking: an embedded script or iframe from, or a link to, a known scheduling or reservation service (for example Calendly, Acuity, Housecall Pro, ServiceTitan, Jobber, Square Appointments, Vagaro, Booksy, OpenTable, Resy, Zocdoc, Setmore, Fresha, Boulevard, Zenoti, Tekmetric and Toast reservations). A business’s own custom booking page would not be detected.
- Chat widget: a script or iframe from a known live-chat or messaging vendor.
- LocalBusiness structured data: a JSON-LD @type of LocalBusiness or any of its subtypes in the current schema.org vocabulary, such as Dentist, Plumber or Restaurant.
- Platform: identified from page code signatures; sites with no recognizable signature are counted as other or custom.
Limitations
What this report cannot tell you.
- Field data was dropped. We planned to report whether each site had real-visitor Core Web Vitals in Google’s Chrome UX Report, via the PageSpeed Insights API. Without an API key the service returned a quota error (HTTP 429) on the first request, so that metric is not in this report. Every speed figure here is a lab measurement.
- One run, one moment. Each site was tested once. Lighthouse scores vary from run to run, sometimes by several points, and sites change. The medians are steadier than any single site’s number, which is one reason we do not publish single-site results.
- The sample is not random. It comes from search results, which favor businesses that already rank. It is 117 sites, not every business in Mesa, and the industry groups are small, so differences between groups should be read loosely.
- Homepage only. A business may have a form, booking or structured data on another page. We measured what a visitor lands on first.
- Detection has blind spots. Booking and chat are detected from a list of known vendors; a custom booking system would be missed. Forms inside third-party iframes count as forms, but their fields could not be checked for labels. Lighthouse could not complete a run on 13 of the 117 sites: 12 servers answered its request with an HTTP 403 refusal and 1 was stopped by a browser security warning. The speed, accessibility, best-practices and SEO figures therefore cover 104 sites. Page checks read each page after it loaded, without scrolling or tapping, so anything that appears only after interaction was not seen.
- Automated accessibility checks are partial. Lighthouse finds a subset of accessibility problems. A good score does not mean a site is fully accessible, and this report makes no legal judgment about any site.
Questions about the report.
Which businesses were tested?
117 independent small businesses with a Mesa, Arizona street address published on their own website, across home services, health, professional services, food, auto repair, salons and retail. Franchises, national chains, dead or parked sites, lead-generation sites and every site Mesa Web Designers has built were excluded. We do not name any business or publish any single site's result.
What does a Lighthouse performance score actually measure?
Lighthouse is Google's open-source testing tool. On a mobile run it loads the page on a simulated mid-range phone over a throttled connection and scores how quickly the main content appears, how long the page is frozen by scripts, and how much the layout jumps around. Scores run 0 to 100; Google labels 90 and up good and below 50 poor.
Is a lab score the same as what real visitors experience?
No. A lab test is one controlled load, which makes sites comparable with each other but is not a measure of any real visitor's experience. Google also publishes field data from real Chrome users, but we could not collect it for this sample, so this report uses lab measurements only and says so.
Why does a click-to-call link matter so much?
On a phone, a tel: link turns the number into a button. Without one, a customer has to memorize or copy your number and switch apps to dial it. It costs nothing to add and is one of the cheapest fixes a small business site can make.
How can I see how my own site compares?
Run your homepage through the free Website Health Check on this site, or through Google's PageSpeed Insights, which uses the same Lighthouse engine. Test on mobile, test more than once, and compare against the medians in this report.
Can I cite or use this research?
Yes. Please cite The 2026 Mesa Small Business Website Report by Mesa Web Designers and link to this page. If you need a figure that is not shown here, ask; we will only share aggregates.
Recognize your site in here?
Get a straight answer about yours.
Tell us the address and what is bothering you. The developer who would do the work reads it and sends back a flat, written number, usually the same business day. No account managers, no mystery quote.
- Flat and in writing before a dollar moves
- Answered same business day, usually faster
- Prefer talking? Call (480) 525-7582
See what your project costs. It takes about a minute.
Rather talk? (480) 525-7582. 0% financing on any project, 25% down. See all pricing
Be the site that beats the median.
Fixed-price websites from $1,500, flat-priced repairs from $149, and one developer who answers the phone.