OK, Google: What is voice search?
This week: Voice search. How does voice search work and how should publishers consider voice queries for news?
Hello, and welcome back. Jessie here. Spring finally, maybe arrived in Toronto this weekend. Soon: the cherry blossoms.
This week: Voice search. As audience editors, our goal is expanding the reach of our journalism. Voice search is a new vertical of potential readership.
Join our Slack community to chat SEO any time.
Let’s get it.
In this issue:
What is voice search?
What does it mean for publishers?
Can publishers optimize for voice search?
What is voice search
A voice search is a verbal query sent to a search engine – often done with a smart speaker such as a Google Home, Amazon Echo or Alexa or another device. It must be a question or query: a verbal command to turn on a podcast or turn off a light does not count as voice search since it does not require the smart assistant to crawl the web for an answer.
Depending on the product you use, the default search engine may vary. Amazon Alexa smart speakers, for example, use Microsoft Bing’s search engine, but Google Home’s smart assistant uses, well, Google. Most mobile phones will default to Google, but some will use Bing.
A user asking, “Hey Alexa, tell me what is happening in Ukraine right now?” is a voice search (playing a news brief from a top-ranking publisher).
Both return content, but only one surfaces news prompted by a question.
If you have a podcast, test to ensure your podcast is returned when from a voice action. (“OK, Google, play The Daily” should return the latest episode from The New York Times.)
Why it matters: According to Search Metrics, almost one-in-three global online users make use of voice search. Ahrefs found that 41 per cent of adults and 55 per cent of teens use voice search, predominantly for quick facts, directions, to find businesses or do research about products.
THE HOW TO
What to know about voice search
Voice search is a series of questions and answers, sent and received verbally. The search engine is able to understand multi-part searches and accurately link together topical queries. For example, if you start by asking, “What time does the Raptors game start tonight?” then ask, “Who are they playing?” Google will understand the “they” to be the Toronto Raptors.
For news search results, Google will provide the name of the publisher, along with instructions telling the listener they can skip to the next story. Google will read aloud the beginning of the article, and send links to that story, and several others, to the listener’s phone. Then Google will read the next story (from another publisher).
Listeners can skip through search results by saying, “Next article,” or by asking another question.
Three things to know about voice search
Voice searches will often be longer queries and pulled from featured snippets, while the answers will vary depending on the device used to ask the question (mobile versus desktop, smart speaker versus phone).
As readers move from typing queries to speaking questions to a smart device, those queries get longer and are more conversational.
In text-based queries, users remove superfluous words like “the,” “of,” or “and” when typing. The result is shorter – but more robotic – queries.
For news, a reader might search on their phone, “ukraine news” for the latest developments in Russia’s war. Through voice search, a reader might ask, “What's happening in the war in Ukraine today?” or “Why was Boris Johnson in Kyiv this weekend?”
For evergreen content, instead of searching “baked oats recipes,” a reader might ask a more precise question, such as: “Find me a vegan baked oats recipe with no banana.”
With voice search, queries are much closer to how people talk in real life.
This changes the strategy from a keyword-based approach to primarily user-intent focused. Instead of targeting a set of key terms, we’re trying to surface more precise content that answers a question – often written conversationally.
A text search on a mobile or desktop returns an array of results – which readers can see and sift through. With voice search, the search engine returns its best match, reads a snippet of the piece and asks a user if they want more information. A search on a mobile or desktop device for a recipe will include multiple recipes (from a variety of publications), while voice search will only return the recipe Google thinks is the best match.
Featured snippets from top-ranked pages
According to Backlinko, 40 per cent of all voice search answers were pulled from a featured snippet and about 75 per cent of all answers come from results ranking in the top three for the query.
The length of featured snippets has been increasing, Moz reports, with “how” queries dominating rich results. To rank in voice search for snippets, focus on longer-tail, natural-sounding keywords – these are the questions readers are seeking out.
Pro tip: If you’re looking for pages to experiment with for voice search, pages that already rank for rich results could be a place to start.
Read more: Review the structured data newsletter for more tips on snagging rich results.
How to optimize for voice search
Data (like Google Trends might provide) does not currently exist for voice searches. Instead, news SEO for voice search is about finding and answering questions that people might be looking for verbally.
Identify questions to answer
Voice queries are longer and more conversational. Use a keyword research tool like Ahrefs or SEMRush to identify long-tail phrase match questions on a topic (i.e., “What are the western conference NBA standings” instead of “nba standings.”).
It’s unnecessary to create a new page to answer a question just for voice search. Instead, if you find a question that is likely to be asked via voice search, see if an existing story answers a similar question and add it there as an H2 heading. Or, create a Frequently Asked Questions (FAQ) or content pillar page for a topic to house the answers to longer, conversational questions, and use the FAQ Schema (more on that next).
Structured data for voice search
Use structured data markup (also known as schema) on your pages. Schema can help your content rank in position zero (the featured snippet, the top spot in organic rankings), and helps drive higher click-through rates.
Level up: Keep an eye on the Speakable Schema from Google, which is currently in beta. This schema “identifies sections within an article or webpage that are best suited for audio playback using text-to-speech.”
Google recommends speakable content that takes about 20-30 seconds to read – about two to three sentences. Content must have concise headlines and/or summaries.
Google uses this data to provide answers on topical news queries on smart speakers. This currently only works for users in the U.S. with Google Home products.
Other tips for voice search optimization:
Formatting your answers: Start by asking your smart assistant the query. What is the content format? If Google is returning a list or short blurb, that’s a clue about how you should structure your content. The correct on-page HTML elements – section or subheadings – will help here, too.
Effort over impact: It is not necessary to create individual pages for each potential voice search. Most results come from pages with more than 2,000 words – so it’s unlikely these pages were specifically designed for voice.
Speed, speed, speed: Make sure your pages are mobile friendly and fast.
Promote and redistribute: Content that is highly shared (on social platforms) with numerous backlinks is more common in voice search.
How much does voice search matter?
It’s unclear how much traffic voice search can generate. Voice search is still considered a niche.
From a verbal query, users get a response (which often answers the question) and the link sent to their phones. A reader needs to follow up on their device to click to your publication. This is a multi-step process that is unlikely to return a flood of audience.
However, it can help build brand awareness for your publication (the name of a publisher is provided in results).
An added benefit: The work of finding and answering reader questions is always a worthwhile audience effort. Creating service journalism that helps readers where they are – and where they consume their news – is valuable work, period.
Having the foundational elements of SEO solidly in place will only help your success in voice search. Start there, expand as resources allow.
The bottom line: As audience editors, our job is to surface our journalism wherever readers might be looking for it – and that can include at home with a smart speaker. While it should not be the priority, it can be a complementary approach to reach a broader audience.
THE JOBS LIST
These are roles across the globe we see that are audience positions in journalism. Want to include a position for promotion? Email us.
The New York Times is hiring an Associate Director, Editorial Search.
Barry Adams: Why are internal links to topic hubs such a powerful SEO tactic for publishers?
YouGov: Where Americans get their news and who they trust for information
Lily Ray: Google’s 2021 and 2022 product reviews updates: What happened?
SEMRush: How did the web change in 2021?
Backlinko: The definitive guide to voice search