Why voice search matters
Voice search is appealing to early adopters because they're offered new features to interact with, and in time it'll be something that's second nature to many. People are starting to get used to giving voice commands to a system to search and find items on their shopping list. Being aware of this trend will help retailers keep up with how their customers are shopping, and get ready for the inevitable moment where voice search is no longer a 'nice to have'.
Blackberry and Nokia didn't switch to touchscreens until much later than their rivals, and ten years down the line they are attempting to revive their businesses as sales have crashed. If you don't board the technology trends train and keep a serious eye on what's going on, there's a very real risk of getting left behind.
Retailers starting to adopt voice search are currently giving consumers a novel experience. Designer fashion site Farfetch's iPhone app, for example, is very easy to use by giving simple instructions like “black dress".
Historically, we are used to searching for products using text search with keywords, but it's more instinctive to do a voice search using natural language and say “I'm interested in red shoes under £50, size 6". Typing long, complicated sentences or word combinations to get to what you're looking for is less appealing.
Existing technology, however, has limitations, often only being able to cope with a few keywords to give meaningful results. And when we speak, we talk naturally. Humans aren't conditioned to speak in a staccato fashion. Think how silly a user would feel saying “Shoes women's red £50 size 6".
When currently searching by typed keywords, users have to 'learn' how the search engine works and adjust their queries to get the results they are looking for; potentially adding or removing words or adjusting the order of words, for example.
They shouldn't have to do this.
Voice search will need to remove this obstacle if it is to become widely adopted. It should enable the user to converse and the search engine should understand what the user means.
This presents retailers with quite an undertaking. The voice search function will need to be pretty sophisticated to accommodate someone speaking naturally in sentences when they're searching for products. The challenge is to understand everybody.
This is where AI comes in. It would need to be able to understand different sentence constructions because we all think differently. We all use different words to describe the same thing, and the system would need to understand all of those combinations and give the same results.
To put it another way, when you do a text search, you go through multiple levels of filtering in your head before typing what you're looking for. You're confident you know what the system expects of you, and you translate it into typed words. When you speak it, you don't have that luxury, because you don't have time to think and say “red shoes", then wait for 10 seconds, then say “for a wedding" and so on.
Essentially, we're asking voice search applications to translate an infinite number of complicated thoughts, and it will likely be one or two years at least for voice search technology to evolve and become widespread.
But it will happen. We're watching, innovating and mastering the role of voice search.
Conversational shopping is where retail is headed.