“Alexa,” I asked my little robot assistant for a test of voice search, “find me a local doctor.”
“Based on your zip code,” she answered, “here are some popular choices.” She rattled off a few options, with one of our clients in first place, and then suggested that I put my address into the Alexa app for greater accuracy.
I was impressed, because in my experience, voice search doesn’t always work that well. Research on the subject finds that most queries of Alexa, Google Home, and other such devices are just commands: “Read me my book,” “Turn the bedroom lights off,” “Turn the volume up.” Actual questions mostly include asking about the weather, traffic, our calendars, and other rudimentary self-centered questions.
We ask Alexa for information pretty frequently at our house, though. She’s good with incontrovertible facts, like the average weight of an ant. At least half the time, however, Alexa answers our questions with, “Sorry, I’m not sure.”
Siri lost my trust through a series of conversations in my car.
“Directions to Shogun.”
“I found several gun shops fairly near you.”
And there was the day Siri took us out into a field in Kansas and said, “You have arrived at your destination” when it was clear that there was no restaurant there. Sunflowers, horses, no restaurant for miles.
And literally not one single person on our team uses voice search on our computers.
We’re not alone. Voice search was less popular in 2018 than in 2017 (when people thought it was fun). But more Americans of all ages use assistants like Siri, Alexa, Cortana, and Okay Google. Now that we’re over asking questions for fun, we are very likely to see an increase in voice search for practical reasons.
For example, it turns out that people are increasingly using voice assistants in the bathroom, where they suddenly want the coffee pot turned on while they’re in the shower or perhaps are inspired to add toilet paper to their shopping lists.
Is voice search for marketing?
As long as we just want to the know the weather or to turn up the heater, voice search isn’t going to be super useful for marketing.
We know that people search for healthcare solutions first by asking about problems, then narrowing down to solutions and finally researching providers of those solutions. I asked Alexa the causes of joint pain, and she read me a paragraph from the first choice at Google (as I discovered when I typed the same query into my laptop). Surely, if I were looking for joint pain solutions, I would not keep asking, “Oh, yeah? Well, Alexa, tell me about gout,” and then “Alexa, tell me about strains,” and so on through the 10 to 12 searches Google tells us people usually make in these cases.
Local searches for a good place for lunch (our Kansas experience notwithstanding) seem about right for voice search marketing use cases.
But that could change.
Optimize for voice search
Let’s say that voice search will continue to be more important over time. How can you optimize to show up well in voice search?
- You need an answer to a question that can be read quickly and easily. Try to answer questions at the beginning of the page or post, and give a concise answer that can be useful without reading the rest of the post. Alexa read us, “WebMD says Many different conditions can lead to painful joints, including osteoarthritis, rheumatoid arthritis, bursitis, gout, strains, sprains, and other injuries” and a brief list of other things. An alternative on the first page of the SERPs started off, “Joint pain may be caused by arthritis, but there are other possible causes which must be considered. Proper treatment depends on identifying …” Clearly, Alexa made the right choice between the two.
- You need an answer the voice assistants can understand. No disrespect intended, but these are machines we’re talking about. They’re not as good with human language as people are. The reading level needs to be fairly low. This is also what the AMA recommends for your patients, so this is not a problem.
- Most voice search results are also at the top of the SERPs. Practically every query we checked showed us the same results on a computer as the voice search answer we were given. When there were differences, they were usually the logical extension of the points made above.
In other words, if you’re making sure your content communicates well with search engines, you won’t need to go back and re-optimize for voice search.