Voice Search SEO
I would start voice search SEO by auditing my content for real questions, not by adding schema or speeding up the site.
What I’d do first
- Audit your top 10 pages and rewrite each intro to answer a clear question in one sentence.
- Build a conversational keyword map targeting long-tail question phrases like 'how do I make Thai green curry'.
- Mark up genuine FAQs with FAQ schema only after writing the answer in plain text.
- Ensure your Google Business Profile is complete and accurate for local voice queries.
- Test mobile page speed under 2.5 seconds on a real 4G connection.
The path I'd take
If I were starting voice search SEO today, I would not begin with schema or speed. I would begin with content that answers a real question. Your page on Thai green curry might rank for "thai green curry recipe" but fail on voice because nobody speaks that phrase. They say "how do I make Thai green curry?" or "what's the easiest Thai curry?". You need to audit your top 10 pages and for each one ask: what question does this page answer? If you cannot answer in one sentence, rewrite the intro.
Next, build a conversational keyword map. Use a tool like AnswerThePublic or the keyword planner to find question phrases. Target [long-tail keywords](/long-tail-keywords/) that have question words: what, how, where, why. For example, instead of targeting "pizza Seattle", target "where can I find a late-night pizza place in Seattle". Then write an answer that lives in the first paragraph after the H1.
Structure that answer clearly. Put the direct answer in bold or in a simple sentence. Voice assistants pull the snippet. You want to win the [featured snippet](/featured-snippets/). Test your page by reading the first 50 words out loud. Does it sound like an answer? If not, rewrite.
Once the content is solid, add FAQ schema for the real questions on the page. Not generic ones. Then optimise mobile speed because most voice searches happen on phones. Run PageSpeed Insights and get your Largest Contentful Paint under 2.5 seconds. I would [use keyword research](/keyword-research/) to find question gaps across your site.
I would prioritise answer clarity over everything else. If you have one page that answers "how to tie a tie", that page will win voice results even if it loads a bit slower than a competitor. But don't take the speed risk.
Watch-outs
The biggest watch-out is optimising only for head terms. Voice queries are longer and more conversational. "Best pizza" is rarely spoken. "What's the best pizza place near me that delivers?" is. If you target only short keywords, you miss the long-tail intent. Use [intent SEO](/intent-seo/) to map each query to a page.
Second, avoid marketing fluff in answers. Voice assistants need a clear, neutral statement. Saying "Our pizza is the most amazing in town" will not get picked. Instead, say "This pizza has a thin crust, mozzarella, and tomato sauce." Neutral and factual. Rewrite every product description that sounds like an ad.
Third, do not add FAQ schema without real FAQs. I see many pages with five FAQ items that are just rephrased headings. That does not help. Write the answer first, then mark it up. Google uses the answer to decide if the page is a good candidate for a voice response.
Fourth, ignore local voice at your peril. A large chunk of voice searches are local: "near me", "directions", "hours". If your Google Business Profile is incomplete or inaccurate, you will lose those queries. Check your listing, update hours, respond to reviews.
Fifth, speed is non-negotiable. Voice users expect instant answers. A 3-second load will make them leave. Test on mobile under 4G. Aim for under 2 seconds. Use tools like Lighthouse. I also watch [SERP features](/serp-features/) to see if competitors are winning voice boxes. Analyse their content structure.
Finally, do not treat voice SEO as a separate channel. It is part of regular SEO. The same principles apply: clear answers, good structure, accurate data. But the emphasis shifts to question phrases and local accuracy. Keep [keyword optimisation](/keyword-optimization/) grounded in natural language.
What I got wrong
I used to stuff FAQ schema onto every page, hoping it would trigger voice results. It didn't. The pages didn't have real, answerable questions – they were just titles rewritten as questions. Now I write the answer first, then wrap it in schema. That alone doubled my featured snippet appearances on one site.
I ignored mobile speed because my desktop score was 95. Then I tested on a Moto G4 with 3G – the page took 5.2 seconds to load. I lost voice traffic because users bounced. I fixed images and deferred JavaScript, getting LCP down to 1.8 seconds. Voice referrals started appearing in analytics after that change.
I also targeted head keywords for too long. I had a page ranking #1 for "running shoes" but it never appeared in voice results because nobody asks "running shoes" – they ask "what running shoes are best for flat feet?". I pivoted to question-based long-tail and saw a 30% increase in organic traffic within three months. I now use [tracking keywords](/tracking-keywords/) to monitor question phrases.
Another mistake: I assumed voice was primarily informational. But many voice queries are transactional. "Order pizza near me" triggers a different intent. I had no page for that. I added a local landing page with a clear call to action and saw phone calls increase.
These mistakes taught me that voice search SEO is not a tactic. It is a discipline of making content answer spoken questions genuinely. Test your own pages by speaking the query out loud and seeing if your page answers it within 10 seconds.
Next step
Quick answers
Does voice search SEO require structured data?
Structured data is not mandatory, but it helps. Marking up genuine FAQs with FAQ schema gives search engines explicit clues. I always write the answer in plain text first, then wrap it in schema. The content wins, not the markup alone. Without real answers, schema is useless.
How can I measure voice search performance?
Measuring voice directly is difficult because Google Search Console does not separate voice queries. I track rankings for question-based keywords, monitor featured snippet ownership, and check appearance in People Also Ask. I also compare mobile traffic trends as a proxy for voice behaviour. Local analytics like 'near me' clicks help.
Should I optimise for voice search separately from regular SEO?
No. Voice search is an extension of regular SEO. The same content quality, structure, and speed matter. What changes is the keyword selection – target question phrases and long-tail queries – and local optimisation. I integrate voice considerations into my standard content briefs. Separate strategies waste effort.
Sources
Primary documentation is linked directly. Anything commercial is marked nofollow.
- Google Search Central — Provides official guidance on structured data and search appearance, foundational for voice-friendly pages.
- Google Business Profile Help — Authoritative reference for local listing accuracy, a key factor in voice search local optimisation.
- Google Search Central: Featured snippets — Explains snippet eligibility, which is directly tied to voice answer selection.
- Search Engine Journal — Provides practical voice search optimisation tactics with real examples.
- Semrush Blog — Offers keyword research workflows that include question-based and long-tail keyword mapping.
- Siteimprove Glossary: Voice Search SEO — Gives a compact definition and audit-oriented framing for a quick reference.
Notes from Callum Bennett.