Voice Search

[Jeffrey]

Hi! Do any of you have a graphic or document showing where each Voice Search device pulls from. For example: Google Home, Alexa, Siri, Cortana. I know I can guess at where they pull from but does anyone have a graphic or anything showing this? Thanks

 

[Mindquest]

@Jeffrey +1 one to that!

I am also trying to understand what boxes to tick off in terms of citations or where to list/claim a business so that it is on Google Home, Alexa, Siri, Cortana...etc.

 

[BipperMedia]

@Jeffrey -- not quite the graphic you were looking for, but here's my 2 cents and hopefully this helps to move the discussion forward:

Google Home: these are pulled most from featured snippets in Google search results and also Google Maps / GMB's for locally targeted searches. You can see / hear a direct correlation between featured snippets and voice results.

Alexa: Amazon (obviously) for product and Amazon shopping related data, a combination of Bing for direct search results and queries, and Yelp (via Apple Maps) for local business related searches.

Siri: a combination of Apple Maps and Yelp, but Siri also relies heavily on Google search results.

Cortana: web search results from Bing and also Bing's Knowledge Graph

All, or most... or all, of these platforms leverage (or look for) Schema markup for different types of content (@types).

But I definitely look forward to other, more "expert knowledge" feedback as this is an excellent question!

 

[Ampere]

How do they get service area businesses (like plumbers/electrician) if SAB's can't list on Apple Maps?

 

[BipperMedia]

How do they get service area businesses (like plumbers/electrician) if SAB's can't list on Apple Maps?

Yelp... Apple Maps... but Apple Maps relies heavily on Yelp... so mostly Yelp

 

[Mindquest]

@Jeffrey -- not quite the graphic you were looking for, but here's my 2 cents and hopefully this helps to move the discussion forward:

Google Home: these are pulled most from featured snippets in Google search results and also Google Maps / GMB's for locally targeted searches. You can see / hear a direct correlation between featured snippets and voice results.

Alexa: Amazon (obviously) for product and Amazon shopping related data, a combination of Bing for direct search results and queries, and Yelp (via Apple Maps) for local business related searches.

Siri: a combination of Apple Maps and Yelp, but Siri also relies heavily on Google search results.

Cortana: web search results from Bing and also Bing's Knowledge Graph

All, or most... or all, of these platforms leverage (or look for) Schema markup for different types of content (@types).

But I definitely look forward to other, more "expert knowledge" feedback as this is an excellent question!

Great info! Found this article which lines up with your thoughts:

How to optimize your local business for voice search - Search Engine Watch

Specific strategies for local businesses trying to understand the potential voice search will have on their bottom lines.

www.searchenginewatch.com

So, where should I cite my business?

Each voice assistant relies on different and sometimes multiple data aggregators for answers to local search queries:

• Siri
  • Search: Google
  • Business listings: Apple maps
  • Reviews: Yelp
• Alexa
  • Search: Bing
  • Business listings: Yelp and more recently Yext
  • Reviews: Yelp
• Google Assistant
  • Search: Google
  • Business listings: Google my Business
  • Reviews: Google my Business
• Cortana
  • Search: Bing
  • Business listings: Bing
  • Reviews: Yelp

 

[BipperMedia]

Great info! Found this article which lines up with your thoughts:

How to optimize your local business for voice search - Search Engine Watch

Specific strategies for local businesses trying to understand the potential voice search will have on their bottom lines.

www.searchenginewatch.com

Nice... that definitely looks more official and elegant than my post

 

[Mindquest]

@BipperMedia Your write up was perfect, it was so hard to find this info until your reply.

The other thing about voice that is hard to pin down (research wise) is related to length of answer. How long should an answer be? Can the answer be too long thus losing consideration for a voice answer? Or will it take a portion of an answer only?

I doubt there is a perfect length (maybe I am wrong) like a title tag or meta description but wondering there are any guidelines people are using. Of course with all the types of voice assistants there are probably specific requirements for each one.

Thoughts anyone?

 

[Ampere]

So if Siri only gets business listings from Apple Maps, how do they find a plumber, which is a service area business that Apple Maps does not allow on it?

 

[BipperMedia]

So if Siri only gets business listings from Apple Maps, how do they find a plumber, which is a service area business that Apple Maps does not allow on it?

Siri: a combination of Apple Maps and Yelp, but Siri also relies heavily on Google search results.

So when it comes to Siri, you have a combination of:

• Yelp
• Apple Maps
• Google search

I would say that for business listing data that is not available in Apple Maps, Siri simply leverages Yelp.

This is also true of Alexa.

I'm thinking that a hidden powerhouse in the world of voice search platforms is Yelp.

I really hate to even say that... because I'm not a fan of Yelp.

They literally blow up our client's phones as soon as we get started with our SEO work.

AND... we have definitive proof that they've presented only 1, 1 star review on a listing before but then when it was verified, multiple 5 star reviews magically appeared -- meaning, it's almost like their platform purposefully blocks certain reviews / ratings from showing until a listing a verified -- and then once verified, that's when they blow up your phone with their army of sales people.

But still... the point holds true -- I think Yelp's data and platform are heavily leveraged in the voice search world.

 

[BipperMedia]

@BipperMedia Your write up was perfect, it was so hard to find this info until your reply.

The other thing about voice that is hard to pin down (research wise) is related to length of answer. How long should an answer be? Can the answer be too long thus losing consideration for a voice answer? Or will it take a portion of an answer only?

I doubt there is a perfect length (maybe I am wrong) like a title tag or meta description but wondering there are any guidelines people are using. Of course with all the types of voice assistants there are probably specific requirements for each one.

Thoughts anyone?

The most common voice responses (i.e. voice search results) are the featured paragraph snippet.

And the optimal length is 45 words long (293 characters) with a maximum length of 95 words (or approx. 750 characters).

If your content doesn't fit the answer perfectly - according to Google's perception of your content - but your page is currently the most authoritative for the voice search response, then Google will create a mashup of your content to present the most comprehensive voice response.

Meaning, Google will grab snippets / sentences from multiple paragraphs and stitch them together to present the best voice response in the voice search result.

In the environments where there's a stitched voice response, this is where the best opportunities exist for you to create a more comprehensive paragraph that's optimized specifically for voice search results.

Here's the optimal structure to follow for that:

<h2> or <h3> subtitle tag presented as a question / or... the question written in bold.

45 words / max 95 words of a direct, to the point answer / response to the question.

Place this toward the top of your page.... or, place it anywhere within the page of content and wrap it in some kind of unique css style to make it stand out.

I hope that was helpful.

 

[Mindquest]

And the optimal length is 45 words long (293 characters) with a maximum length of 95 words (or approx. 750 characters).

thanks for all the great info on this!!

Do you have a resource on where you got the character range for voice? Obviously it will depend (Siri vs Alexa...etc), just trying to understand the justification for this best practice.

Thanks!

 

[Tim Colling]

I'm thinking that a hidden powerhouse in the world of voice search platforms is Yelp.

I really hate to even say that... because I'm not a fan of Yelp.

Yelp is very frustrating. There is a new indie movie out now about Yelp, called "The Billion Dollar Bully". It's a bit over the top, but it provides an accurate illustration of how aggravating they are for small businesses.

 

[Ryanl]

The most common voice responses (i.e. voice search results) are the featured paragraph snippet.

And the optimal length is 45 words long (293 characters) with a maximum length of 95 words (or approx. 750 characters).

If your content doesn't fit the answer perfectly - according to Google's perception of your content - but your page is currently the most authoritative for the voice search response, then Google will create a mashup of your content to present the most comprehensive voice response.

Meaning, Google will grab snippets / sentences from multiple paragraphs and stitch them together to present the best voice response in the voice search result.

In the environments where there's a stitched voice response, this is where the best opportunities exist for you to create a more comprehensive paragraph that's optimized specifically for voice search results.

Here's the optimal structure to follow for that:

<h2> or <h3> subtitle tag presented as a question / or... the question written in bold.

45 words / max 95 words of a direct, to the point answer / response to the question.

Place this toward the top of your page.... or, place it anywhere within the page of content and wrap it in some kind of unique css style to make it stand out.

I hope that was helpful.

I just started checking out your website. You're in Athens! I'm in Gwinnett.

 

[BipperMedia]

Yelp is very frustrating. There is a new indie movie out now about Yelp, called "The Billion Dollar Bully". It's a bit over the top, but it provides an accurate illustration of how aggravating they are for small businesses.

Yes... we unfortunately have made it a standard practice to warn our clients... once we start doing what we do, Yelp will definitely be blowing up their phones and they can pretty much take anything a Yelp rep would say as "ultimately a sales tactic".

 

[Tim Colling]

once we start doing what we do, Yelp will definitely be blowing up their phones and they can pretty much take anything a Yelp rep would say as "ultimately a sales tactic".

^^Truth

 

[Tim Colling]

Every once in a while, I accept one of their calls, just to see what they are using as high-pressure sales tactics. It's painful but informative.

I used to do custom database work for a timeshare marketing company back when I had a small custom software agency (back in the 1980s). I thought that they were outrageously high-pressure back then, but they can't hold a candle to the tactics of Yelp telesales people now.

Ugh.

 

[Ampere]

I guess Yelp just doesn't like me, or maybe my voice carries my brute manliness over the phone so well that they were scared to call back

When they called, I told them straight up that I would not do any business with someone cold calling me and not to call again. No more calls for a period of time until 1 day after I got a positive review (one of many) and they mentioned that review and asked if I wanted to use any of their programs. This time I told them that I would not give them a penny because they hid many of my positive reviews from people with long Yelp histories, but showed the 1 negative review I ever received which was from someone with no other Yelp history at all. They said they can't do anything about that, so I told them not to call me again. They haven't.

 

[BipperMedia]

I don't really want to get on too much of a Yelp bashing spree here... the point is, Yelp is heavily integrated into voice platforms and they are a presence in the local search market - that's just reality.

We use Yelp as a citation distribution source > get our client's data straight on their profile > get their listings verified when necessary... and then we move on (with warnings to our clients to be prepared for the calls).

 

[pony]

@BipperMedia, do you have any decent ways of analyzing what types of searches are being performed by voice so you can market towards it?

Or, do you just try to optimize for snippets in hopes that it will also cover voice searches? I don't like the feeling of the unknown.