Connect with Machines to Connect With People
Connect with Machines to Connect With People
Teodora Petkova frames the challenge for writers in the new world of voice search.
Voice Search and the New Writing Order
To speak or not to speak to machines?
A new writing order might sound a bit too broad a perspective to use when thinking about voice search but it is not. :)
In this webinar hosted by +David Amerland and +SEMrush's and with awesome presentation by +Zara Altair (her slides are available in the link here: https://www.semrush.com/webinars/voice-search-meet-the-one-answer-challenge/) we had great time and got wonderful questions covering all kinds of concerns website owners have when it comes to ranking. One of them was by +Monika Schmidt and somehow resonated with my recent context of writing about writing :) for people and machines.
So, Monika asked:
Question: “How do you tell the search machine, which snippet out of your content should be taken for talking back as search result in voice?”
To meet the one answer challenge for Monika’s question, I would go for :
Short answer: You use data-ish :) [best Linked Data-ish (ref. SEO, Linked Data and Your Fridge Browsing the Web - http://www.teodorapetkova.com/poiesis-of-relationships/what-linked-data-has-to-do-with-seo/)]
Long answer: See below :)
Let’s deconstruct the question to take all the opportunities if offers for deeper understanding:
How do you tell a machine something
Well, you use all kinds of code, the computer code languages being so many and different, just like our own. And speaking to machines takes all kinds of forms - from giving instructions, struggling to speak their own language - think Press button A and then button X and then hit twice this and this to the sophisticated ways today we can give instructions to machines: Find all the places near me that offer cappuccino with almond milk.
When talking to search engines (and other systems) we can now use natural language and the system does its best to extract meaning and intent from our words and the context we are asking these words in. What we can do, as content creators to "tell the machine" something, is we can add pieces of code, as metadata, which describe what we have said.
“Being specific about the machine” It is good to think in terms of the specific cases we want to talk to a machine: that could be a search engine, a voice app, a chatbot, a personal assistant, a cognitive system that manages our content. And this breadth of things we want to make our content clear to, gives me the opportunity to tell you: speak Linked Data :) In the SEO world it is mostly known as RDFa. In Plain English RDF is just a three-part sentence: Fox Cafe sells Almond Milk. Fox Cafe is located in Sofia. (for the curious, get ready and dive deeper here: https://schema.org/docs/datamodel.html)
How do you instruct a machine which snippet of your content you want it to read
You don't :) unless you know exactly how this machine works, what is the specific context and what is the exact outcome needed. In all other cases, in which we often happen to be, we use mark-p with common designators, that is with additional machine-readable information. (see for example Zara's 26th slide on speakable schema)
Search results in voice
How is Google, or any other (cognitive) system behind an assitant (say the Bing behind Cortana, or the Watson behind Siri) rendering results, that is applying the principles of information retrieval and knowledge discovery.
Short Answer - highly connected data.
This answer ties well to +Andrea Volpini's comment: "the debate short vs long content - extremely relevant for voice” and Paul Meyers' question: What tools are available for testing?.
And the last question is key to the long answer about voice search results: we need to test. Although we (me particuliarly) are include to think interconnectedness is something so natural, it takes a lot of technological developments so that it becomes part of our everyday life. And the people form +WordLift are doing a lot of experiments with voice search so you should check them out.
And with that - have a great Saturday and welcome to the beautiful and very challenging new writing order, where we need to be able to define very well our content goals and the ways we will bring machine-intelligence behind the content intelligence we are aiming for.
Originally shared by Anton Shulke
Zara Altair David Amerland Teodora Petkova
https://www.youtube.com/watch?v=RFDlU1pQP_w
Teodora Petkova frames the challenge for writers in the new world of voice search.
Voice Search and the New Writing Order
To speak or not to speak to machines?
A new writing order might sound a bit too broad a perspective to use when thinking about voice search but it is not. :)
In this webinar hosted by +David Amerland and +SEMrush's and with awesome presentation by +Zara Altair (her slides are available in the link here: https://www.semrush.com/webinars/voice-search-meet-the-one-answer-challenge/) we had great time and got wonderful questions covering all kinds of concerns website owners have when it comes to ranking. One of them was by +Monika Schmidt and somehow resonated with my recent context of writing about writing :) for people and machines.
So, Monika asked:
Question: “How do you tell the search machine, which snippet out of your content should be taken for talking back as search result in voice?”
To meet the one answer challenge for Monika’s question, I would go for :
Short answer: You use data-ish :) [best Linked Data-ish (ref. SEO, Linked Data and Your Fridge Browsing the Web - http://www.teodorapetkova.com/poiesis-of-relationships/what-linked-data-has-to-do-with-seo/)]
Long answer: See below :)
Let’s deconstruct the question to take all the opportunities if offers for deeper understanding:
How do you tell a machine something
Well, you use all kinds of code, the computer code languages being so many and different, just like our own. And speaking to machines takes all kinds of forms - from giving instructions, struggling to speak their own language - think Press button A and then button X and then hit twice this and this to the sophisticated ways today we can give instructions to machines: Find all the places near me that offer cappuccino with almond milk.
When talking to search engines (and other systems) we can now use natural language and the system does its best to extract meaning and intent from our words and the context we are asking these words in. What we can do, as content creators to "tell the machine" something, is we can add pieces of code, as metadata, which describe what we have said.
“Being specific about the machine” It is good to think in terms of the specific cases we want to talk to a machine: that could be a search engine, a voice app, a chatbot, a personal assistant, a cognitive system that manages our content. And this breadth of things we want to make our content clear to, gives me the opportunity to tell you: speak Linked Data :) In the SEO world it is mostly known as RDFa. In Plain English RDF is just a three-part sentence: Fox Cafe sells Almond Milk. Fox Cafe is located in Sofia. (for the curious, get ready and dive deeper here: https://schema.org/docs/datamodel.html)
How do you instruct a machine which snippet of your content you want it to read
You don't :) unless you know exactly how this machine works, what is the specific context and what is the exact outcome needed. In all other cases, in which we often happen to be, we use mark-p with common designators, that is with additional machine-readable information. (see for example Zara's 26th slide on speakable schema)
Search results in voice
How is Google, or any other (cognitive) system behind an assitant (say the Bing behind Cortana, or the Watson behind Siri) rendering results, that is applying the principles of information retrieval and knowledge discovery.
Short Answer - highly connected data.
This answer ties well to +Andrea Volpini's comment: "the debate short vs long content - extremely relevant for voice” and Paul Meyers' question: What tools are available for testing?.
And the last question is key to the long answer about voice search results: we need to test. Although we (me particuliarly) are include to think interconnectedness is something so natural, it takes a lot of technological developments so that it becomes part of our everyday life. And the people form +WordLift are doing a lot of experiments with voice search so you should check them out.
And with that - have a great Saturday and welcome to the beautiful and very challenging new writing order, where we need to be able to define very well our content goals and the ways we will bring machine-intelligence behind the content intelligence we are aiming for.
Originally shared by Anton Shulke
Zara Altair David Amerland Teodora Petkova
https://www.youtube.com/watch?v=RFDlU1pQP_w
Thanks for sharing Zara Altair ! And it is ironic that I totally forgot that the G+ system won't allow the share of a reshare :) I mean the text in the reshare...
ReplyDeleteTeodora Petkova I did, too, for a moment. Then I saw Google not sharing your words. :)
ReplyDeleteZara Altair looks like only the first person who post d gets reshared. Good to know :) I miss the Ripples feature.
ReplyDeleteHi
ReplyDeleteThanks
ReplyDeletekeramatzade.com - گروه تحقیقاتی و آموزشی کرامت زاده| ثروت |عشق|موفقیت|قرآن|اعتماد بنفس|
http://modirebimeh.ir/
https://keramatzadeh.com/