Voice search and your business: does "hey Siri" find you?

Voice search is not a separate channel that needs its own strategy. It is the local search your customers already do, phrased like a question, with one big difference: the assistant reads out a single answer instead of showing a page of options.

That difference should bother you more than any adoption stat. When results are a list, being third still gets you seen. When the result is one name spoken out loud, you are either the name or you are nothing. Most local websites are still written for the typed keyword era, and most of the advice about fixing that is either recycled hype or a package someone wants to sell you. Here is what actually changes, and the short list of things worth doing.

How is a spoken search different from a typed one?

People type in keywords and speak in questions. Nobody says "plumber etobicoke" to their phone; they say "who's the best plumber near me that's open right now?" Spoken queries are longer, conversational, and loaded with qualifiers that tell you exactly what the customer needs: open now, near me, takes new patients, tonight.

The answers change shape too. When Backlinko analyzed 10,000 Google Home results, the average voice answer was just 29 words, written at a ninth-grade reading level. Twenty-nine words. That is the entire window your business gets. If no page on your site answers a customer question in two plain sentences, there is nothing for the assistant to read, and it moves on to whoever wrote one.

Does voice search actually matter, or was it hype?

Both, and it helps to separate them. You have probably seen the claim that "50% of all searches will be voice by 2020." That never happened. It was never even a real stat: the source was a 2014 prediction by Baidu's chief scientist about voice and image search combined, in China, which got misattributed to comScore and repeated for a decade. A lot of "voice search optimization" packages were built on that misquote.

The behavior underneath the hype is real, though. 35% of Americans own a smart speaker, about 101 million people (Edison Research, 2025), and eMarketer projects US voice assistant users growing from 139.8 million in 2022 to 168.2 million by 2029. On the local side, BrightLocal found 58% of consumers had used voice search to find local business information, and 46% of those users do it daily.

Notice what the device data says: voice search is mostly a phone behavior, not a speaker behavior. 56% of local voice searches happen on a smartphone. The person asking is out, busy, or has their hands full, which is exactly the customer who calls rather than browses.

One honest caveat. The best local voice data is from 2018, and nobody has run a study that thorough since. I read that as a signal, not a gap: voice stopped being a headline and became a habit, the way nobody publishes studies about whether people use GPS. You do not optimize for habits by chasing trend reports. You optimize by being the clearest answer available.

Where do voice assistants get their answers?

Two places, depending on the question. For "find me a business" queries, assistants read from business listings: your Google Business Profile, your Apple Maps listing, and the hours, categories, and reviews attached to them. For question-style queries, they pull from the pages already winning regular search. Nothing exotic. No secret voice index.

The numbers make the second part concrete. In Backlinko's study, 40.7% of voice answers came straight from a featured snippet, and 74.9% came from pages ranking in the top three for the query. The winning pages were also fast and secure: 70.4% were on HTTPS, and they loaded about 52% faster than the average page.

Read those three bars as one sentence: voice does not create a new ranking game, it raises the stakes on the one you are already playing. If you rank fifth for "how much does furnace repair cost," a screen still shows you. A speaker does not. The pages that win voice answers are the pages that already won regular search, structured so a machine can lift a clean answer out of them.

Why does your Google Business Profile do most of the work?

Because for the highest-intent voice searches, the assistant never touches your website. "Find a plumber near me" is answered entirely from listing data: business name, primary category, rating, distance, open or closed, phone number. Siri and Google Assistant read those fields out loud. If your hours are wrong, the assistant says you are closed and offers the next name on the list. You lost a customer without ever knowing they existed.

And these searchers act. In BrightLocal's study, 28% of consumers called the business after a voice search, 27% visited the website, and 19% walked into the location. Nearly three quarters took a real step. That is not browsing behavior; that is buying behavior.

So the unglamorous checklist is the one that pays: exact hours including holidays, the right primary category, a phone number someone actually answers, and a name, address, and phone that match everywhere they appear. Reviews matter here too, because "best plumber near me" leans on ratings to pick the one name to read. If you have not checked the profile in a while, audit it in 15 minutes before you touch anything else.

How do you write content that answers spoken questions?

Structure every important page around the questions customers actually ask, and answer each one immediately. The pattern is simple. Heading phrased as the question. First two sentences give the direct answer, aiming near that 29-word window. Detail and proof below for the humans who keep reading.

Keep the language plain, too. Backlinko found voice answers written at a ninth-grade reading level, pulled from pages averaging 2,312 words. That combination matters: Google lifts a short passage, but it lifts it from a substantial page. Thin FAQ pages with one-line answers do not carry the authority; long pages that bury the answer in paragraph six do not surrender it cleanly. You want depth with the answer on top.

• Write headings as questions. "How much does duct cleaning cost in Toronto?" beats "Our Pricing Philosophy."
• Answer in the first two sentences. Then explain, qualify, and prove it underneath.
• Use the words customers say, not industry words. People ask about a leaking water heater, not "tank failure remediation."
• Cover the qualifiers. Open now, service area, emergency availability, and prices are what spoken queries ask about.

You are reading the format right now. This article opens with a direct answer, runs question-shaped headings, and keeps the language plain. The same structure wins featured snippets, and featured snippets are where 40.7% of voice answers come from.

What is the overlap with AI assistants?

Voice search was the trailer. AI assistants are the movie. The behavior voice trained people into, asking a full question out loud and trusting a single spoken answer, is now how millions interact with ChatGPT and its voice mode every day. OpenAI's own usage research found 49% of ChatGPT messages are people asking questions, not assigning tasks. Asking is the dominant behavior, and local "who should I hire" questions are part of it.

Here is the useful part: the work is the same. Accurate structured business data, real reviews, and pages that answer questions directly are exactly what AI engines cite when they recommend a business. I have covered how AI search is changing local visibility and answer engine optimization separately; the short version is that voice search, featured snippets, and AI citations all reward the same clarity. One effort, three surfaces.

Which makes the buying decision easy when someone pitches you voice-specific services:

Frequently asked questions

It rewards clarity, same as everything else

Every search shift of the last few years points the same direction. Featured snippets rewarded direct answers. Voice raised the price of vagueness by reading out one name. AI assistants raised it again by citing whoever explains things best. The businesses that win all three are not running three strategies; they keep their data accurate and answer questions like a person who wants to be understood.

So skip the voice package and do the boring version: check your profile, fix your hours, and rebuild your key pages around the questions customers actually say out loud. That is standard work in every local SEO engagement we run, and the free audit will show you in 30 seconds how readable your site is to the machines doing the talking.