Spanish Language Website for Texas Service Business: A Three-Gate Decision Framework

A Spanish language website for Texas service business owners is not a branding decision. It is a routing decision with three gates, and in my experience most shops fail the second one without ever knowing it. I have watched contractors spend a month arguing about whether to translate their site, publish a machine-translated mirror of all forty pages, and then send every Spanish caller to a receptionist who apologizes and hangs up.

So let's skip the debate and build the decision instead. Three gates, in order. If you fail a gate, you stop. If you pass all three, the build is smaller and more specific than you think.

Gate 1: Does Your Texas Service Area Actually Speak Spanish at Home?

Not Texas. Not your metro. The ZIP codes you actually dispatch trucks to.

According to the U.S. Census Bureau, more than a third of Texans age five and older speak a language other than English at home, and Spanish accounts for the overwhelming majority of that. But statewide averages are useless for a business that runs a forty-mile radius. Webb County and Hidalgo County are not Collin County. A plumber working the Rio Grande Valley and a plumber working north Fort Worth are answering different phones.

Here is the twenty-minute version of the research, and you can do it today:

1. Pull your last 100 completed jobs out of your field software, whether that is ServiceTitan, Housecall Pro, Jobber, or a spreadsheet, and export the ZIP codes. Those are your real service area, not the map on your About page.
2. Go to data.census.gov and open table S1601, Language Spoken at Home. Filter the geography to ZIP Code Tabulation Area and enter your top eight to ten job ZIPs one at a time.
3. Write down two numbers per ZIP: the percentage who speak Spanish at home, and the percentage who speak English less than "very well."

The second number is the one that decides this. Plenty of bilingual Texans speak Spanish at home and search in English without thinking about it. The households that change your website are the ones where English is genuinely hard, where someone is going to type plomero cerca de mí or reparación de aire acondicionado because that is the language the problem lives in.

The thresholds I use

• Under about 10% limited-English across your top ZIPs: stop here. The demand is not there, and you will spend attention you need elsewhere. Fixing the conversion leaks on your existing pages will produce more booked jobs than a second language will.
• Roughly 10% to 25%: build the minimum version described below. Two pages, hand-written, done properly.
• Above 25%: Spanish is not a secondary track. It is a co-equal one, and the question changes from "should we" to "which language gets the default homepage."

Then run one more check that no keyword tool will give you: ask your dispatcher how many calls in the last year opened in Spanish or got handed to the one tech who translates. If the answer is "a few a week" and your census numbers are thin, trust the phone log. It is first-party evidence and it beats the estimate.

Curious what a site like this costs you in missed leads? See real builds for businesses like yours.

Gate 2: Can Your Business Answer the Phone in Spanish?

This is the gate that kills the project, and it should kill it.

A Spanish page that ranks generates a Spanish call. If that call reaches someone who cannot take it, you have not gained a lead. You have spent money to create a bad first impression and a one-star review in a language your competitors can read and you cannot. In the trades the website is a phone-ringing device, and a ringing phone nobody can answer is worse than a silent one.

So before a single word gets written, answer this: who takes the call? There are only four honest answers, and three of them are workable. Someone in the office is a native or fluent speaker and is on the schedule during your published hours. Your answering service offers a genuine bilingual queue, not a voicemail box labeled "Spanish." You publish a separate tracking number that rings a specific bilingual person's cell. Or nobody, in which case you stop and fix staffing first.

If you pass with a bilingual answering service, test it before you launch. Call your own published number at 7:40 on a Tuesday morning and at 8:15 on a Saturday night and speak only Spanish. Whatever happens on those two calls is what happens to every lead the pages produce.

The same logic runs downstream. If the Spanish page promises a free estimate, the estimator who shows up needs to be able to walk a homeowner through a failing evaporator coil or a split drain line in Spanish. If the tech can diagnose but not sell, the job dies in the driveway. And if your invoices, financing paperwork, and warranty terms are English-only, say so plainly on the page rather than discovering it at signature time.

Gate 3: Which Services Deserve a Spanish Page First?

You are not translating a site. You are choosing two jobs.

Sort your service list by two criteria: urgency and homeowner decision-making. Urgent problems get searched by whoever is standing in the water, in whatever language that person thinks in. Considered purchases get searched by the household member who handles paperwork, which is often the more English-comfortable one. That single distinction tells you where Spanish pays first.

• HVAC: emergency A/C repair beats system replacement. In August, el aire no enfría is a now-problem. A 16 SEER2 changeout is a family conversation held over three days.
• Plumbing: drain and sewer stoppages and water heater failure beat repipes and remodels. Drenaje tapado and calentador de agua are the phrases that ring phones.
• Electrical: panel and outage work beats generator installs and EV chargers.
• Roofing: storm and leak repair beats full replacement, though replacement leads are worth more, which makes roofing the one trade where I would argue for putting the higher-ticket page second rather than skipping it.
• Cleaning, landscaping, auto repair: lead with the recurring-service page, because the buying decision is fast and the relationship is long.

Pick two. The emergency page and the single highest-volume repair page. That is the entire first phase.

Why Does Machine-Translating Your Whole Website Backfire in Search?

Because Google has a written position on exactly this, and because the output reads wrong to the people you are trying to reach.

Google's spam policies specifically name translating content into many languages with automated tools and publishing it without human review as a form of scaled content abuse. That is not a gray area or a ranking penalty rumor. It is a documented policy, and a plugin that spins up forty auto-translated pages overnight is a textbook description of the behavior it names.

The quality problem is worse than the policy problem. Machine translation fails hardest on exactly the phrases that carry your offer. "Free estimate" comes back as estimación libre, free as in liberty, not free as in no charge. "Licensed and insured" comes back as licenciados y asegurados, which in Spanish reads as holding a university degree, not a trade license. Short button text loses its context entirely, so "Book now" turns into a sentence about a book. A homeowner in Pasadena or Mission reads that in about three seconds and decides you are not a real local company.

Then there is regional vocabulary, which no general-purpose translator handles. Texas Spanish follows Mexican usage. A plumber is a plomero, not a fontanero, because the second is Spain and it sounds like a call center in another hemisphere. A water heater is a calentador de agua or, colloquially, a bóiler. Depending on the county, air conditioning is aire acondicionado or simply clima. A free estimate is a presupuesto gratis, not the English-shaped calque. These are not stylistic preferences. They are the words people actually type.

The third failure is structural. Browser-plugin and JavaScript-widget translators, the little flag dropdown in the corner, do not create separate URLs. Google has nothing distinct to crawl, index, or rank, so you get zero search benefit for the trouble. If the Spanish version does not live at its own address, it does not exist as far as search is concerned.

The Minimum Viable Build: Two Spanish Pages, Done Right

Here is the specification I would hand a developer. It is short on purpose.

• Real URLs in a subfolder. /es/reparacion-de-aire-acondicionado/, not ?lang=es and not a widget overlay. A subfolder keeps the pages on your existing domain and inherits its authority, which is the right trade for a local business with one site.
• Self-referencing canonical tags. This is the single most common technical mistake I see: the Spanish page carries a canonical pointing back at the English page, which tells Google the Spanish page is a duplicate and should be ignored. Each Spanish page points at itself.
• Correct language attribute and hreflang. Set <html lang="es">, and pair each page with its English counterpart using reciprocal hreflang="es-us" and hreflang="en-us" annotations, plus an x-default. Both directions, or the annotation is ignored.
• A visible language switcher, no automatic redirect. Do not bounce visitors based on IP address or browser language. Bilingual Texans switch on purpose, and Googlebot crawls from U.S. addresses, so auto-redirection can hide the pages from the crawler entirely.
• Written, not translated. Hire a bilingual writer who knows the trade, or have your best bilingual tech talk through the page in Spanish on a voice memo and let a writer shape it. The page should sound like a person from your county, because the person reading it is. This is the same principle that governs who writes the content for the rest of your site. Proximity to the actual work beats fluency in the abstract.
• Its own phone number. Use a distinct tracking number that routes to whoever passed Gate 2. This is how you measure the whole experiment, and it costs you nothing but a line in your call-tracking account.
• A form that submits in Spanish. Labels, placeholders, validation errors, and the thank-you page. Half-translated forms leak leads at the error state, which is precisely where a nervous first-time customer abandons.

Two pages built to that spec will out-earn a forty-page auto-translated mirror, and they take a fraction of the time. At Turnkey Web we treat the Spanish track as a scoped addition to a build rather than a checkbox, for exactly this reason: the work is in the writing and the routing, not the page count.

What about Google Business Profile and paid search?

Your Google Business Profile stays singular. One real-world business, one profile, one name as it appears on the truck, and creating a second profile for a Spanish version is a guidelines violation and a fast way to get both suspended. What you can do is encourage Spanish-speaking customers to leave reviews in Spanish, which is a genuine local ranking and trust signal in the exact language you are courting. The service area settings on your profile matter more to your visibility than anything on the language side, so get those right first.

On paid search, one detail catches almost everyone: Google Ads language targeting keys off the user's Google interface language and browsing behavior, not the language of the query they typed. Spanish keywords alone will still serve mostly to English-interface users. Change the campaign language setting as well, and point those ads at the Spanish page rather than the English homepage, because sending paid traffic to a homepage wastes it in any language.

What the Law Expects Once You Sell in Spanish

Two Texas-specific items that a general web designer will never raise with you.

First, licensing disclosure. TDLR rules require licensed air conditioning and electrical contractors to include their license number in advertising, and plumbing came under TDLR's authority on September 1, 2021 under SB 621. A Spanish service page is advertising. Your TACLA, TECL, or RMP number belongs on it, in the same prominence as on the English page. Roofing is the exception, since Texas does not license roofers at the state level, which means a Spanish roofing page has to carry manufacturer certifications and insurance documentation to do the same trust work.

Second, and more consequential: under Texas Business & Commerce Code Chapter 601, the three-day cancellation notice in a home-solicitation sale must be in the same language principally used in the oral sales presentation. If your estimator sits at a kitchen table and sells a system replacement in Spanish, the cancellation notice has to be in Spanish. Marketing in Spanish leads directly to selling in Spanish, and selling in Spanish carries a paperwork obligation. Get the document translated properly before the first appointment, not after the first complaint.

How Do You Know the Spanish Pages Are Working?

Give it ninety days and look at four things, in this order.

1. Calls to the Spanish tracking number. Raw count. If it is zero after ninety days of indexed pages, your Gate 1 research was wrong or the pages are not ranking. Check indexing before you conclude anything about demand.
2. Booked rate on those calls versus your English booked rate. A materially lower rate points at Gate 2. The page did its job; the handoff did not.
3. Search Console impressions filtered to the /es/ path. Impressions without clicks means your Spanish title tags and meta descriptions are the weak link, which is usually a sign they were translated rather than written.
4. Average ticket. This is the one owners skip. If Spanish leads book but at a lower average ticket, the gap is almost always in the estimate conversation, not the website.

Ninety days, four numbers. Then you expand to a third and fourth page, or you stop and put the attention somewhere with better returns. Either outcome is a win, because you will have decided it with evidence instead of opinion.

The Decision Tree, in One Pass

Compressed, so you can run it in five minutes:

• If limited-English households are under 10% of your top job ZIPs, then stop. Improve the English site.
• If the demand is there but nobody can answer a Spanish call during business hours, then stop. Solve staffing or bilingual answering first, then come back.
• If both gates pass, then build exactly two hand-written Spanish pages: your emergency service and your highest-volume repair, each at its own /es/ URL with a self-referencing canonical, reciprocal hreflang, your license number, and a dedicated tracking number.
• If those two pages produce booked jobs at a comparable rate within ninety days, then expand to your next two services and translate your estimate paperwork.
• If you are ever tempted to auto-translate the whole site in an afternoon, then reread Google's scaled content abuse policy and don't.

The framework is deliberately conservative because the failure mode is expensive and invisible. A bad Spanish page does not announce itself. It just quietly takes calls you cannot serve and converts warm prospects into people who tell their neighbors you were not ready.

Frequently asked questions

Will Spanish pages hurt my English rankings?

No, provided each language version lives at its own URL with a self-referencing canonical tag and correct hreflang annotations. Problems come from duplication, not from bilingualism. Properly annotated separate pages do not compete with each other.

Should I use a Google Translate widget instead of building real pages?

No. Client-side translation widgets produce no separate URL for Google to index, so you get no search visibility, and the output carries vocabulary errors that make the page read as foreign to a Texas homeowner.

Do I need a separate Google Business Profile for the Spanish version?

No, and creating one risks suspension of both listings. One real-world business gets one profile with the legal business name. Instead, ask Spanish-speaking customers to leave reviews in Spanish.

What if my service area is bilingual rather than Spanish-dominant?

Build the two-page minimum and let the data settle it. Bilingual markets search in English for considered purchases and in Spanish for emergencies, which is why the emergency page comes first.

Can I just have my bilingual tech translate the pages?

Often yes, and it beats a translation service that does not know the trade. Have them talk through the page in Spanish on a voice memo, then hand the recording to a writer who shapes it into page copy.

How many Spanish pages should a Texas service business eventually have?

Match the English site's core service pages plus a Spanish contact page, typically four to eight for most trades. Blog content and About pages rarely justify translation.