As I was attending Pubcon Austin 2016 I made the wise decision to attend a session being presented by Purna Virji. For those of you who don’t know her, she’s the Senior PPC Training Manager at Bing Ads and she’s freakin’ smart. But I digress.
Her session was about tips for shopping campaigns and she talked about voice search. As an exercise she had the room pull out their phones and do a voice search for “show me Yves Saint Laurent bags”. The results around the room were predictably random and she had a screenshot of her result, translated to “Show me you send the wrong bags.”
For the purposes of replication I just did a voice search on my phone for “Yves Saint Laurent” and got this result:
For the record, I don’t speak French so the pronunciation was probably poor. But that’s kind of the point. Voice search is being done by people all over the world. You’ve got different accents in Boston and Texas and Minnesota. And we’re even speaking the same language.
Advantages of Voice Search
Before this starts to sound all doom and gloom, let’s remember some things that voice search gets right:
- Corrects misspellings – Many people don’t have a chance at spelling certain words right. However, search engines have learned a lot about misspellings in their work (how many times have you seen that “Did you mean X” message on a search you did?) and can compensate on tough to spell searches.
- More natural – Queries are done conversationally so you usually see longer queries and that means more intent signals.
- Easier – This is bigger than it seems. Based on research from eMarketer, Display will pass Search in ad dollars this year. It’s not because search isn’t growing, but because display is growing faster. The engines want to see more searches done so they can make more ad revenue. Making search easier and more accessible (my phone is literally within 10 feet of me during 23.5 out of 24 hours) means more ad clicks and more revenue.
Disadvantages of Voice Search
As pointed out by Purna, a big problem has to do with languages, regional dialects and accents. The algorithms will continue to improve, but when the same person could be switching between different languages from one search to the next it has the potential to turn into a game of Mad Gab. Or think about how a can of Coca-Cola could be called “pop”, “soda” or just “a Coke” depending on where in the US you are. This makes your language settings, geographic targeting, and negative keywords that much more important.
Voice search is growing. So much that by 2020 it’s predicted that half of all searches will be done by voice. So you’d better start thinking about longer search queries that are more conversational and use natural language (that means slang, regional dialects and accents).
What are you doing to prepare for voice search? Are you seeing these queries in your accounts already? Share with us in the comments!