Actions on Google “Usage” filters — learning from 2 years of Word Chain, a Google Assistant game
Hello Google Actions and voice developers! I recently discovered some cool features in Actions on Google console, under the “Usage” section. Check out the video I below exploring filters that allow to get more in-depth data of voice app usage.
By playing around with these filters I discovered few things that I thought might be helpful to other developers who consider building an app for Google Assistant (and think about ways to monetise it too).
Observations
Looking at the analytics data of my game called Word Chain (which celebrates 2 years since release by the way!) I noticed few interesting things:
1. Up until November 2019 the traffic was more or less stable and relatively high (3k — 5k daily users). There were spikes in the weekends, which can be explained easily as Word Chain is a game — and players have more time to play games in the weekend.
2. Something’s changed in November 2019 and ever since the traffic never got back to the same level and became less stable.
2. Since April this year the traffic oscillates between 80–150 daily users, with spikes between 3k and 10k users, taking place roughly every 12 days.
3. Vast majority of traffic comes from mobile devices. Surprisingly smart speakers and smart displays represent a very small percentage of users.
4. Roughly speaking, for every user coming from the US (second most frequent visitors) there’s at least 10 users from India.
Takeaways
Reviews like this really open my eyes. Since January this year, Word Chain has a monetisation gateway, which frankly wasn’t optimised for that kind of usage. The conversation design also could benefit from a revamp knowing the localisation of users. Now, having discovered the Surface filters, I’m considering adding extra features to the game — perhaps using Interactive Canvas, to make the game experience more interesting for those using smartphones.
I’d love to find out more about my game’s players: in what situations do they reach for Word Chain? Is it in the comfort of their homes? Or perhaps while driving a car? Do they play on their own or with others around them? What is their age? Knowing answers to these questions would allow to build a lot better conversations and effectively — better experiences.
The unknown
My instant reaction to seeing the stats was “Wow… But, why?”
I should have mentioned that I haven’t advertised Word Chain anywhere other than in occasional posts on LinkedIn, Twitter or Reddit. All traffic is somehow steered and controlled by Google Assistant itself. From what I can tell by looking at the usage stats, Google is actively seeking the best solution to discoverability issues of Actions on Google. It’s a big topic in this very dynamic environment of voice assistants. For now, as it usually happens in other areas of digital advertising (such as SEO), developers can only follow the best practices, but the outcome is never guaranteed.
One sure thing is that if you’re not in the voice app space now, you may miss out the day when the groundbreaking changes to the ecosystem start delivering tangible benefits to the creators of voice apps.
Show your stats
If you also love Actions on Google, and have developed an app that gained some traction with users, I’d love to see the usage stats of your app. Post them in comments or get in touch — perhaps together we can work out some better conclusions and that will lead to building better and more engaging voice apps.
Share it
If you liked this article, don’t forget to clap once 👏 or twice 👏👏 and share it with your friends. Thank you for your support!
