We're a few months into having both a Google Home air freshener and an Amazon Echo hockey puck, and it's been interesting to see which one of the two we seem to chat to most.
As it stands, Alexa is our most popular artificial friend. Not because the service is better, but simply because "Alexa" is easier to say than "OK, Google". I know Google is probably better. On the occasions I've tried more complex questions on both devices, Google's years of natural language and search algorithms expertise shine through over Amazon's, but I just cannot get used to saying that awful mouthful of a trigger phrase.
I've been wondering what it is about the trigger phrases that make one easier to say other another. When I even think about saying "OK, Google", my tongue feels strained - the clash of that K next to those two G sounds. I suspected it might qualify as a tongue twister, so I asked a linguist friend what she thought.
She explained to me that linguists would describe the K and the G sounds as "minimally different", in that they share most of the same properties. They come from the same position in the mouth, making them hard to say in succession, much like the tongue-twisters described in this article "Why tongue-twisters are hard to say". "OK, Google" is hard to say in the same way that "She sells sea-shells" is difficult.
I wasn't imagining things - "OK Google" is simply much harder to say than "Alexa". Performing lingual gymnastics every time we want to turn on a light just isn't going to fly.
I noticed while watching some of the Google I/O coverage that many of the presenters talking on the topic of the Google Home/Assistant were in fact saying "Hey, Google" instead of "OK, Google". I assume because "Hey" sounds more natural and also it's easier to say. I know most of the I/O stage-talks are heavily scripted and show pre-recorded Assistant conversations rather than live interactions, but I just shouted "Hey, Google" and our Google Home does respond to that (in fact, it responded every time I played the video I linked to). The alteration in the phrase is not a documented customisation, rather they're making the "OK Google" phrasing a little more forgiving in certain cases.
Not all words are going to be easy for all people to say - some sounds are impossible for some people to make, depending on which language you learned in your youngest years, or due to speech impediment. Sometimes it's even more banal - I wanted to use the Amazon trigger-word "Computer", since with my British accent I imagined Picarding it up around the house, but sadly it only understood me when I feigned a terrible American accent. Disappointing.
It seems right that customisation in this area is the way to go if these voice-activated systems are going to be equally accessible, but Google currently offers no official way to change their trigger phrase. Amazon allows a choice of 3 trigger words, each different sounding from the other - Echo, Alexa and Computer. Just this small selection opens up the opportunity to find a verbal trigger word that is comfortable to say and easy to remember, not just for one user in the home but for a whole family. It's not fully customisable, but it's a good start.
The story on the grapevine is that Google doesn't allow customisation of the trigger word because they don't want to personify and name their assistant - they don't want you naming your device after a person and then potentially treating it badly (the latter probably being a fair concern). You'll note that 2 of the 3 Amazon words are not what you'd typically consider person-names, so this excuse is a little weak and doesn't explain not offering a choice at all. Plus the Google Assistant literally already sounds like a human being, rather than a robot / synthesised voice, and you're encouraged to interact with it via normal human-to-human conversational patterns.
There are unofficial workarounds for android phone users (for google assistant), but they won't help you with the assistant on your Google Home. If you can't say "google" easily, or something that sounds remarkably similar, you're gonna have a frustrating time.
Voice controlling our entire homes could be really amazing, particularly for those with mobility or vision issues, but they're no good if they don't work for real people, their real voices and their real vocabularies. Google and company should do the hard work to make it simple for as many as possible and I hope that's the trend we'll see.