Google vs. Amazon – End game.

Assistants are too stupid for mass adoption.

• While growth in the digital assistant space is slowing, the battle looks like it is almost over although Google still has one or two tricks up its sleeve.
• The latest report from Voicebot.ai (see here and here) echoes the theme from CES 2020 which clearly pointed to a maturation of the digital assistant market.
• Data from Amazon shows that skill growth has slowed markedly during 2019 despite the fact that the 100,000 barrier was broken in September 2019.
• In 2019, 13,936 new skills were introduced compared to 31,009 in 2018 and 18,731.
• I think that there are two main reasons for this.
  • First, hobbyists: Amazon has a very experimental approach to life and as such has been happy for almost anything to be uploaded onto its platform.
  • This meant that anyone with some coding knowledge could create a skill for Alexa which led to a large number of very low-quality skills being created.
  • The excitement of voice as a computing platform is now beginning to wane and as a result, the hobbyists have lost interest causing the number of skills to fall.
  • At the same time, this should lead to an increase in the quality of skills created which is an area where Amazon has badly lagged Google to date.
  • Second, stupidity: Digital assistant are badly suited to be powered by the kind of AI that is in widespread use today.
  • This AI is based on deep learning and as such only really works well where the task being addressed is very well defined and where nothing changes.
  • Human conversations are neither of these things which is why all digital assistants very often fail to fulfil the requests being made.
  • I think that this is the biggest barrier to both the wider adoption of digital assistants and further increases in their use in the digital lives of users.
  • The result is likely to be a maturing market while there is no radical improvement in AI to make the assistants work better.
  • RFM research indicates that Alexa is worse than Google Assistant, Baidu Duer OS and potentially Yandex’s Alice.
  • However, even the leaders are far too stupid to be of any real day to day use leaving being mostly used to tell the time, set timers and play music.
• Despite these limitations and its clear inferiority, Amazon continues to dominate the digital assistant in the home space where Google has not really been able to close the gap.
• Apart from the smart home hall at CES 2020 where Google had clearly helped exhibitors pay for their stands and where the white-suited bods where everywhere, the show itself felt like that the battle between Amazon and Google was over.
• Google still has a trick up its sleeve which is in mobile phones where its position as the default Android assistant has led to its dominance of assistant usage on Android.
• Google has now gone one step further which is to enable the assistant on feature phones that run KaiOS in emerging markets which has had a significant impact.
• The best example of this is in India where Google Assistant “skills” known as actions available in Hindi are now in second place at 7,554 compared to 18,826 in English.
• This goes some way to cementing Google’s position in India which is a market that Amazon is fighting tooth and nail with Instacart to win.
• The net result is that a two-player market in most of the world looks set to stay with Amazon being the dominant player and China and Russia doing their own thing.
• This is negative for Google which had a shot at overturning Amazon with its superior product but its inability to really execute in smart speakers has led it having to settle for second place.
• I remain pretty indifferent to Google, but I would prefer it to Amazon whose valuation I still cannot stomach.