By Adam Stetzer
In the movie Her, Joaquin Pheonix plays a lonely and heartbroken man who develops strong romantic feelings for his mobile operating system. “Samantha,” as he calls it, speaks to him, listens to him, and ultimately becomes a major part of his life.
As technology stands right now, we may not be at the point of making real “human” connections with our mobile devices, but we do talk to them. We talk to them a lot.
Mobile technology is progressing quickly, and soon, speech recognition software will become even more advanced and far more prevalent than it is now.
Ultimately, improved speech recognition software will have a significant impact on how the public uses search engines. As search engine queries evolve, the results will naturally follow suit.
Consequently, SEO practices will be affected and businesses will have to develop new techniques (or alter current methods) for optimizing their websites.
Speech recognition software: the road to success
Smartphone speech recognition software has been frustratingly slow, glitchy, and spotty at best. At its worst, the burgeoning technology has produced some embarrassing and problematic communication blunders.
Basically, talking to Siri has been somewhat like trying to converse with your older grandmother after she’s taken out her hearing aids for the night.
It’s a far cry from the convenient hands-free technology creators originally set out to build. A new experiment, though, has suggested that there is a solution on the horizon.
Putting new technology to the test
James Landay, a professor of computer science at Stanford and co-author of this recent study said:
“Speech recognition is something that’s been promised to us for decades, but it has never worked very well. But we were noticing that in the past two to three years, speech recognition was actually improving a lot, benefiting from big data and deep learning to train its neural networks to produce faster, more accurate results. So we decided to formally test it against humans.”
Landay and his research team, which included scientists from Stanford, the University of Washington, and Baidu Inc., took Baidu’s Deep Speech 2 software and challenged 32 texters to a competition of speed and accuracy.
To test Baidu’s reportedly advanced speech recognition software, the researchers recruited a group of 19- to 32-year-olds to type messages on the Apple iPhone’s built-in keyboard in a race against the machine.
“They grew up texting, so we’re putting speech recognition up against people who are really good at this task,” explained Landay.
The participants either typed or verbalized about 100 everyday phrases. Some subjects performed the task in English while the others did so in their native Mandarin. Regardless of the language, the results were clear – Baidu’s Deep Speech 2 software was significantly faster and more accurate.
In English, speech recognition was an astonishing three times faster than typing, and the error rate was 20.4% lower. In Mandarin, the speech recognition software performed with a staggering 63.4% lower error rate.
Stanford computer science PhD student and co-author Sherry Ruan said:
“We knew speech recognition is pretty good, so we expected it to be faster, but we were really actually quite surprised to find that it was almost three times faster than typing on a keyboard.”
The future of speech recognition software
Now that researchers have been able to quantify the success of speech recognition software, they hope it will inspire engineers to design better user interfaces that take advantage of this technology.
“We should put speech in more applications than just typing an email or text message,” expressed Landay. “You could imagine an interface where you use speech to start and then it switches to a graphical interface that you can touch and control with your finger.”
Though software company Baidu does not currently plan to make Deep Speech 2 available to the public, they are integrating it into their own apps in China.
Regardless, it seems that improved speech recognition software is on its way to smartphones and other mobile devices here in the United States. Before we can discuss how this technology will impact user behavior and affect the way we do business, we must first answer the question, ”Who is using speech recognition software and why?”
It may not come as a surprise, but the majority of people using mobile personal assistants like Siri fall into the generational category of ‘millennials‘. The breakdown looks like this:
- Ages 18-29: 71%
- Ages 30-43: 59%
- Ages 44-53: 39%
- Ages 54+: 38%
Now that we’ve established the “who,” it’s time to talk about the “why.” Why are we using speech recognition software? What’s the purpose?
According to a Northstar Mobile Voice study, teens and adults use their phones’ speech feature for different reasons.
The majority of teenagers (43%) use speech recognition to make a phone call, while 38% ask for directions and 31% use the search feature to get help with homework. On the other hand, 40% of adults use it to ask for directions, while 39% dictate texts and 31% use it to call someone.
Many people are starting to use speech recognition for search engine queries. It is this application in particular that will impact businesses, digital advertisers, and SEO specialists.
How will voice search affect SEO?
There are bound to be differences between how we type our search queries and how we verbalize them. As user behavior changes, so, too, will search engine results. The question is will you be ready to adapt to these changes in order to keep your website optimized?
Here are the two major ways voice search will affect SEO.
Query length and phrasing
When we type our search queries, we rarely input more than two words. We typically don’t write full sentences, knowing that Google will get the gist.
While right now, voice searches are still relatively short, with most capping at around three words, it is important to note that voice search is still in its infancy.
As mentioned previously, voice recognition software hasn’t historically been all that useful; only now are we starting to see more advanced programs that can actually comprehend and accurately communicate with us.
As the software improves and we become more comfortable using it, it is highly likely that voice search queries will start to resemble natural language.
Natural language demonstrates intent more strongly. When we speak in full sentences, our meaning becomes clearer to our audience. It’s pretty simple, right?
Here’s an example:
The user types “air conditioner” into the search bar. Google doesn’t know if they want to buy one, repair one, learn their history, or see images of them. Now, if they were to use voice search and say “Where can I buy an air conditioner?” then Google can provide them with more appropriate results.
For online marketers and SEO specialists, knowing users’ intent can be extremely helpful. They can strive to rank on question phrases with a higher likelihood of action.
Modifying your SEO strategy
As Internet user behavior changes, your business will need to respond and adapt.
Here are a few ways you can take action to maintain your website’s search engine optimization.
- Add relevant question phrases to your keyword list (i.e. How much does it cost to travel to Los Angeles?)
- Identify your most valuable question phrases and avoid questions that don’t suggest action (i.e. What time is it in Los Angeles?)
- Include filler words in your keyword questions, like “the”, “to”, “I”, and “for.”
Mobile technology has been progressing not in fluid strides but rather in giant leaps. It’s exciting, it’s impressive, but it’s also exhausting. Businesses need to be on their toes, ready to react and adapt at a moment’s notice.
Speech recognition software may have been around for awhile now, but it has yet to play a significant role in the shaping of our modern society. The time has come. Is your business ready?
Source:: Search Engine Watch RSS