Google has announced a new markup targeting eligible publishers who want to have their content within Google Assistant. The new markup, called speakable allows publishers to mark up those parts of an article which are important to be read aloud by Google Assistant and devices such as Google Home. The features of the markup are available for English language users who are based in the US. Google stated that it will launch capability in other languages and nations once a significant publishers implement speakable.
At present, Google lists the markup as “BETA” along with a disclaimer which reads that it’s “subject to change”. Google is working on this feature and users might come across changes in either guidelines or requirements. The specification presently is on Schema.org which denotes that the speakable markup signifies sections of a web page which are specifically ‘speakable’ and are highlighted as being suitable for text-to-speech conversation. Other parts of a web page can also be usefully spoken in specific circumstances. The ‘speakable’ property pinpoints the parts which have a high chance of being used for speech.
The technical guidelines include:-
- Instead of highlighting a whole article with speakable structured data, concentrate on the major points. This enables listeners to acquire an idea of the story rather than having TTS readout do away with important details.
- Avoid adding speakable structured data to content which can be confusing in voice-only and voice-forward situations, such as photo captions, source attributions, and datelines(location from where the story was reported).
The content guidelines include:-
- Content which has been specified by speakable structured data should have crisp headlines and/or summaries which offer users comprehensive, helpful information.
- For best user experiences on audio, Google suggests around 20-30 seconds of content per section of speakable structured data or around two to three sentences.
- In case you include the top section of the story in speakable structured data, Google suggests that you rewrite it to break up the details into individual sentences to ensure that it reads more clearly for TTS.
Google’s example of how this might work with a Google Home Device has been shown above. But Google stated:-
When people ask the Google Assistant – “Hey Google, what’s the latest news on NASA?”, the Google Assistant answers with an extract from a news article and the name of the news organization. Then it asks if the user would want to hear another news article and also sends relevant links to the user’s mobile.