Joaquin Phoenix, as Theodore Twombly, speaking to Samantha, “the first artificially intelligent operating system . . . a consciousness that knows you.” From the 2013 movie “Her.”
The world Spike Jonze envisioned in the 2013 movie “Her” no longer seems that far off with the buzz around Artificial Intelligence (AI), Internet of Things (IoT) and the exploding growth of voice search and voice-activated devices.
By the end of 2017, according to eMarketer, 35.6 million Americans will have used a voice-activated device at least once per month.
And that’s just digital assistants on stand-alone voice-enabled speakers like Amazon’s Echo and Google’s Home.
Once you factor in the voice assistants on smartphones, such as Apple’s Siri, Google’s Assistant & Voice, and then include Microsoft’s device-agnostic Cortana, 60.5 million Americans will have used voice-activated search at least once per month over the last year.
With this huge shift, Tank has been working with our clients to ensure their websites are optimized for voice search. It’s pushed us to rethink content creation and how we approach SEO and Paid Media strategies.
Matching User Intent to the Right Content
It’s only been in the last few years, since Google’s release of Hummingbird, that the machines have made significant advances toward understanding user intent. Now, with a greater emphasis on semantic search, Google is better able to infer what you want from your search queries.
So let’s put it to a test using an iPhone 8 Plus, running iOS 11.2.
I’m currently traveling on business and won’t be home for a week. I need to pick up a couple of new shirts and I wear Patagonia as my travel uniform. I pull up the Google Voice app and ask, “Where is the nearest Patagonia store?”
No surprise here, Google pulls up the closest store listing – but here is where it gets interesting. I’m going to follow-up with a question that, as a stand-alone query, is too vague to provide a useful result.
In the search above, I simply asked, “What time does it open?” with no mention of Patagonia. Since I had already established that I was looking for a Patagonia store, Google assumed I wanted to know what time that particular store opened.
So far, so good. It’s only when I start with a different question that Google attempts to redirect me to another location.
I ask Google, “Where is the nearest place to buy Patagonia?” and receive a complete different response.
This time Google announces through my iPhone speaker, “Coats and jackets are available for $199 at REI, 2.1 miles away. They are also available starting at $59 at three other stores in that area.”
REI has been able to dominate the page position on mobile by delivering paid advertisements through Google Shopping. The ads are compelling with the use of ratings and, by using their geo-location, I know they are only 2.1 miles away. The organic listing for the nearest Patagonia store has been pushed off the viewable screen area, with Patagonia’s Store Locator moving up to the first position.
REI’s listings aren’t perfect. They served up two women’s Patagonia coats and my intent was to look for men’s shirts. However, it was enough to make me want to click through to see if I could save some money and, since I am an REI member, I can get money back on any purchase at the end of the year.
For comparison purposes, I wanted to see how Siri treated semantic search.
Just like Google, Siri was able to tell me where the nearest Patagonia store was and the opening time based on the question, “What time does it open?” However, when I asked Siri, “Where is the nearest place to buy Patagonia?” she returned a very different result. This time Siri directed me to the nearest Patagonia store, which was 3.1 miles away, even though REI was closer at 2.1 miles away.
Voice Search Increases Complexity for Marketers
Being able to understand all of these aspects for SEO, Paid Media and on-site Content and how they interact together is a complex issue. Without a complete view into the entire customer experience and the ability to impact change in each area, marketers will be unable to respond appropriately to a world where 50% of searches are expected to be voice-activated by 2020, according to Comscore.
At Tank, we forecast that voice search will become increasingly complex in coming years as users grow tired of just asking simple questions and AI continues to improve, allowing for a more conversational approach that may allow you to tell your virtual assistant several descriptors before a search is actually completed.
Imagine instead if I could just say, “Google, what’s the lowest price on a men’s Patagonia Sweater in gray in a size large? I’m willing to drive 30 minutes after work today.”
That may sound far-fetched at present, but the technology already exists and soon will be assembled together.
Optimizing Your Website for Voice Search
As noted, voice searches are more conversational and natural and are typically much longer that we see in text keyword searches.
10 years ago search, marketers would start with a list of the top keywords by volume and then add in a laundry list of long tail terms to find incremental volume.
Voice search completely blows up that model. Long tail is the strategy. Marketers must begin thinking about user intent and how to drive those conversations.
Tips for Optimizing Your Website for Voice Search
- What are the types of questions that your prospects or customers ask when they call your company? Start compiling a list of those frequently asked questions – don’t just paraphrase, document them verbatim. The goal is to have a document that reflects how real people speak and ask these questions.
- After the list is complete, create pages within your website that match these FAQs. Group similar questions on the same page and make sure that all of the questions and the answers you provide remain conversational.
- If you have a site search function on your website, consider applying semantic search thinking to it or risk looking antiquated.
- Ensure your technical team is up-to-speed on how to use structure data markup and can apply the correct schemas to provide voice searches with richer information. This will have the added benefit of helping text searches by giving your content proper context.
- Apply semantic thinking across all Search programs. Don’t forget to add these long tail voice phrases to paid programs to ensure you are reaching people at each state of intent.
Tony Roy is the head of digital marketing and eCommerce strategy at Tank. He has generated over $1B in incremental eCommerce revenue during his career. Tony created the first health social platform and the first mobile package-tracking app. He has been recognized by the Smithsonian Institute for innovation as the product manager for FedEx’s Internet shipping application.