Thinking Aloud: Optimizing for Voice Search in 2020
What is Voice Search?
Voice Search is speech recognition technology that allows users to search on websites, search engines, or apps by speaking to a device – such as a mobile phone, computer, or smart speaker. This post will go over a few different types of voice search and how to optimize for these.
Echo / Eyes-free / Desktop & Mobile Voice
When it comes to voice search, it’s important to nail down exactly what type of voice search is being referred to. Whether that be Amazon Echo searches, which can be categorized as general searches and product searches, “eyes-free voice” search, which denotes searches that involve devices that have very small screens or no screens at all like a Google Home, or a mobile or desktop voice search. Each of these categories has a bit of a different approach when it comes to optimizations.
Amazon Product Search Optimization
When a user searches for a product on their Echo, Echo Dot, Echo Show, or Echo Spot, Alexa recommends a single product to the user and asks them if they want to buy it. Assuming the user has no past purchases in this category, in which case Amazon will recommend the past purchased product, specific product searches will return Amazon’s Choice product that is relevant to that specific query. Amazon Choice products are selected by Amazon around the following parameters:
- Rating & Reviews
- Shipping Speed
- & Other Factors
If there is no Amazon’s Choice item available for that search, then the highest organically ranking Prime offer will be selected. This includes both fulfilled by Amazon and seller-fulfilled Prime products.
Worth noting: Amazon uses conversion rate (conversions / user page views) to inform the organic ranking of products
Eyes-free Voice: Insights from Google Guideline
“An eyes-free voice assistant is an electronic device that can understand voice queries from a user, and give audio responses or take actions on the user’s behalf. For example, Google Home is a smart speaker that lets you interact eyes-free with the Google Assistant. An eyes-free device does not have a screen, but may be connected to other devices in the user’s home. For example, an eyes-free device may be able to connect to the TV to play a video. For this task, the term eyes-free is used for devices that have no screens or keyboards (or very limited screens and keyboards), so the primary method of interacting with the device is talk to it, and listen to audio responses.”
– Evaluation of Search Speech, Guidelines
Google rates “eyes-free voice” answers & actions around 4 total factors that fall into 2 groups.
- Needs Met Rating – How complete is the answer? Does it fully satisfy the true & implied intent of the query. (This is my terminology, Google does not explicitly state this, but does use this example.)
- i.e. “Who is the president of the united states?”
- true – “the president of the united states is the elected head of state of the United States”
- implied – “The current president of the United States is Donald Trump.”
- Speech Quality Rating
- Length – Length of answer. Concise is best, but they do consider if additional context or information could improve the answer.
- Formulation – Grammar & Syntax. Does it seem to represent the way a native speaker would answer the question? Is the source clear?
- Elocution – How accurate are the pronunciations? Was the intonation natural? Was the speed appropriate?
Some of these can be optimized for while others cannot. Keep in mind both the implied and true intent of a query and try to answer both. We can shape the text to be more succinct, but we’re not going to change the cadence in which Alexa, Siri, or Google speak. We can optimize the text in a manner that would be friendly if read aloud. Avoid stringing together numbers that make for confusing responses.
General Voice Search
Optimizing for general voice search is very similar to optimizing for SEO best practices in general. Below are some general notes on how to optimize for general voice search.
- Page Speed is critical.
- HTTPS is common.
- Concise is best.
- Strong domains are more likely to show.
- 9th grade reading level average.
- Title tag doesn’t seem to matter.
- Generally pulled from long-form content.
- 75% of voice answers are from top 3 organic positions.
Beyond these points, there are a few other areas worth focusing in on to ensure the best chance to show for any voice query, location, Google My Business, and Frequently Asked Questions.
Location is extremely important in voice searches. Voice searches are much more likely (3x according to early 2017 data) to be done on a mobile device than a desktop device. This means that Google is extremely likely to know the exact location of the searcher, so localized queries will be highly location dependent. This rolls well into the next point of Google My Business.
Google My Business
Google My Business listings are extremely important as many local searches are for a business or entity. Setting up and claiming the GMB listing will help ensure each business can show as a result of a voice search.
Frequently Asked Questions
Frequently Asked Questions & Answer the Public are your friends. Most voice queries are structured similarly to how a FAQ section of a website will structure their content. Ensure the site has a FAQ section somewhere and utilize Answer the Public to expand on existing questions to supply better topic coverage to Google.