Exploring the Past, Present and Future of Voice Recognition

  • twit

Posted: 10/04/2017

artificial_intelligence_004

Leaps and bounds for voice recognition

The first voice recognition software appeared in the 1950’s and had the capabilities of understanding digits. A decade on and another system called the Shoebox system came to light which could recognise 16 words was created by IBM. Quite an achievement for 1962 with computers now considered archaic. It would take another two decades in development for the speech portfolio to reach a few thousand recognised words and another decade until most of the population reached familiarity with voice recognition (through the rather unfortunate rise of voice activated menus).

All of this however, a far cry from the intelligent systems which are now at the forefront of our daily lives today. And what factor can we attribute to our relationship with artificial intelligence? Mobile devices have certainly created a huge impact on how we retrieve information and consumers are becoming more demanding on getting information quickly, accurately and efficiently. On average a person can speak 110 – 150 words per minute as opposed to the average typing rate of 38-40 words per minute. Furthermore voice activation is no longer a one way conversation; with the rise of giants producing at home devices such as Google home and Amazon echo alongside voice assistants such as Apple’s Siri and Amazon’s Alexa which not only retrieve results but also manage to intelligently decipher certain queries and at times with added humour, it’s no wonder that we are now seeing 1 out of ten searches as a voice query.

Voice recognition a passing fad?

Clearly there has been a fundamental shift in how we relate to technology, particularly in the recent years. With major advances towards voice recognition the industry is set to be worth a staggering $3 billion by 2020

In a recent trend report called Speak Easy, from JWT and Mindshare More than 60 percent of all participants indicated they’d use the technology “all the time” if voice assistants could interact “as well as a human can.” And it’s not just the younger generation benefiting from voice assistants. Although the 18-34’s dominate the market, the older generation well into the 50’s are also benefiting from its ease of use.

Admittedly there are still stumbling blocks. From understanding the user intent, (see video below for an example of where dialects and use of slang can be problematic) and the certain amount of specifism required to return an accurate result there is still room for improvement but we can be fairly certain voice assistants are not a passing trend.   

Why featured snippets are so important for voice recognition

A featured snippet is the information which is shown at the top of the page above the listed results. The snippet selected by Google, displays content from within one of the pages ranking on page one that directly answers the question searched for without the user having to visit the actual page. This can appear in the form of a paragraph, list (as shown below) or table format.

blog_fs

If Google can find the answer to a question presented via voice query, it will usually present information from the featured content.

As we know the featured snippet is designed to give the user the most clear and comprehensive answer to their query which makes them the perfect juice to power voice search. You may have already created search worthy content but having the content that appeals to conversational search is another ball game.

There is no way to create a featured snippet for your brand – Google chooses what to feature but there are things that you can do to enhance your chances.

How do we optimize for voice activated search?

So, what can we learn from a First and foremost the user behaviour is different. The way users address their query is conversational rather than based on keywords and broken sentences. With more long tail queries and questions the intent is clearer. Therefore creating content to mirror this behaviour is imperative for marketers.

For example, a typed search query may be similar but the conversational query is likely to reveal the actual intent.

Example of a typed search query:

‘Boxing gloves, Everlast’

Conversational query:

‘Hi Siri, are Everlast boxing gloves made from real leather?’

So, how do we optimize content for search? By predicting likely queries around your product/service and providing comprehensive and easily digestible information structured around the brand.

First of all, does the brand have content that answers the who, what, where, when, how, typical queries covering all levels of intent?

For example, a shop that sells boxing gloves might want to think about content that answers the following queries based around research and informational searches.

  1. What size boxing gloves should I buy as a beginner?
  2. Where can i find boxing gloves for children?
  3. How to wrap my hands for boxing
  4. What is the best boxing brand?
  5. What to consider when buying boxing gloves
  6. What are boxing gloves made of?
  7. Why are good quality boxing gloves  important?

A good place to start is by using a tool from answerthepublic.com. This website visualises auto results from Google and Bing based around a topic of interest.

 

Capture

 

Creating comprehensive content is basically giving your audience what they need by answering as many potential questions in one go. If you can create this and make it engaging and attractive to Google you have the potential for top spot worthy content.

When creating a content strategy consider the following to add weight, deph and value:

  1. Is the content engaging to searchers
  2. Is the content itself satisfactory? – Well researched and answering the query explicitly?  
  3. Is the content diverse and unique?
  4. Is the  raw content quality good (what pleases the ‘Google spiders’)
  5. Is the content fulfilling the comprehensive criteria (presented in various attractive formats)?

Depending on the nature of your business and the competition already creating content, you might have to think outside of the box. If your competitors are already dominating the landscape across likely search queries, looking into the more niche queries or paying more attention to the value of the actual content might give you the advantage. This could include more diverse outputs and customised formats – graphs, interactive elements, comprehensive guides and data driven sources for knowledge.

Additionally we must remember to keep analysing and adjusting. Featured snippets are based on answering queries based on intent. Understanding intent of potential customers will also give us an idea on what content to create. You can read more about optimisation and understanding intent

Finally, make sure you have created a digital footprint by allowing your brand to have visibility. Including your band in directories and optimise your content for NAPS with accurate details and additional useful information such as open hours on the top business listings such as Apple Maps, Google My Business, and Bing Places for Business.

All in all, this should give you the first steps to not only increase your visibility in the search space but to improve the brand experience for your consumers by providing value added content.

 

More Great Digital Articles

hiring

We’re Hiring: Search Marketing Consultant

Posted by Tom Etherington - November 01, 2017

VIEW FULL STORY   >