John Campbell, ROAST’s Head of SEO, presented findings from our Voice Search Vertical Reports at the Smart Voice Summit in London. If you missed the event, or if you were there but would like a reminder, here are five key takeaways from his presentation:

 

Answer rates vary across verticals

The ROAST team queried 10,000 key phrases across 20 verticals and found that the disparity between these verticals was huge. Thy found that the lowest answer rates, and therefore the greatest voice opportunities, were in Transport, Property and Technology.

 

Answer boxes don’t always equate to assistant results

Getting answer boxes is an important part of optimising for assistant answers, but having one doesn’t mean you automatically get the other – as is always the way in SEO, there are multiple factors at play. For example, in the Insurance vertical 26% of assistant answers did not match the answer box found from the web.

 

Update Google My Business 

Assistants pull data for location-based queries from Google My Business, so restaurants and hotels need to make sure theirs have the correct information (phone number, address etc.), and as many good ratings as possible.

 

Steer clear of ‘whitelisted’ phrases

‘Whitelisted’ phrases are queries for which Google has pre-defined conversation paths. For example, if you were a flight comparison site there would be no point in you trying to gain voice results, because the assistant will always choose to take potential customers into its own service – Google Flights. Picking your battles is vital in the voice search arena.

 

Implicit Invocation Optimisation

It’s a bit of a mouthful, but Implicit Invocation Optimisation, the practice of trying to ensure that the assistant automatically chooses to take consumers into your action, is incredibly important. If done successfully it enables brands to communicate with consumers, without the consumers having to remember what the specific action is called.

 

Download the Whitepaper to find out more about The Variance for Verticals in Voice Search