How Voice Search May (Might) Influence Local Search

Submitted by annsmarty on Mon, 2013-01-21 12:50

Local marketing

Alexis Gui talks about the Impact of Siri on Local Search and provides some stats and food for though on how local search may be influenced by voice search technologies:

If voice search technology becomes more robust and reaches a point where it offers high user value by making search simple, integrated, and hyperlocal, it could have a dramatic impact on how we choose to perform local searches.

Recaps:

1. People tend to use voice for local search (i.e. voice searches are “three times more likely to be looking for local results” than searches on Google’s browser site or text-entered queries on the Google iPhone app")

2. Google voice search is faster (peobably because "Siri searches for answers using direct data sources before performing a web search") while Siri is better for certain popular search categories like movies or restaurants by "presenting more relevant results in a cleaner format" (more here)

3. Tips for marketers:

Siri pulls relevant results from directories like Yelp and Localeze. Therefore, data consistency across directories and search engines is crucial
The impact of reviews is multiplied when it comes to voice search as Siri pulls ratings and reviews from various sources like Yelp and Google Places and use them to rate results
Keywords! Voice queries are still search queries. BUT it makes sense to consider the difference between natural language and written search terms.

Have you ever come across any research ref difference between natural language and written search terms?

Comments

Long tail

Submitted by annsmarty on Mon, 2013-01-21 17:21.

There's an interesting comment on this from Duncan Johnson:

The first thing that came to my mind was the last thing mentioned in the article. KEYWORDS - it'll be interesting how they will differ from typed search terms. The traditional long tail terms get a little shorter when people switch from desktop to mobile. I predict the opposite will happen as voice recognition becomes more sophisticated. I think the length if the long tail will more than double.

More comments

Submitted by annsmarty on Tue, 2013-01-22 11:52.

From G+ Thread:

Misspellings won't matter, since the instructions are spoken.
Ambiguous terms may need qualifying with more information. For example, "pool new york" ... are you looking for swimming or billiards?
Finally, there is a huge difference in the way that people use spoken language (and I don't mean accent, but actual syntax). Education, upbringing, ethnicity, age etc. will produce different spoken instructions, and therefore different pools of search results.
Can a computer detect Irony, Sarcasm, or even localisation of accents. Here is the UK we have so many different accents, Welsh people and Geordies might as well speak Swahili to each other.

Seek Relevance

Submitted by Bill_Scully on Wed, 2013-01-23 05:35.

For some reason I always speak in a more natural sentence structure in voice search than I've trained myself to type in a search box. Example "what times does..." vs. "Store name, location, times". So if the engines can deliver the right results, I predict longer natural sentences will dominate voice search.