We should look at how many startups from YC die within 4 years and make the comparison. I might be wrong, but the ratio could be very similar.
Still, Stadia will fail for sure ^^
It doesn't mean that zero people use it. Thousands or millions of people used Hangouts, Google Reader, Google Code, or Picasa - they'd have been considered big YC successes.
If big YC companies all got shut down if they failed to turn into another letter in the FAANG acronym, maybe people would become reluctant to use services from YC companies.
I still use some Google services, but I check out my data on a regular basis, I'm not buying a Fitbit, and I'm not subscribing to Stadia because I want those things to last longer than I think Google will keep them around.
Other than google reader I doubt any of these would have gotten traction at all without big g behind them.
So if Google is going to launch something under the Google brand they either have to be happier with things that aren't making 10s of billions, or not make so many efforts and burning their users/customers/devs at every step. Why as someone downstream do I want to dedicate my time and money to their product when they've shown that unless it becomes a massive success it will be folded?
Google Code: GitHub sold to Microsoft for a good chunk of change. GitLab, Bitbucket, Source Hut, all seem to be doing fine.
Hangouts: Instagram, Snapchat, Facebook Messenger, WhatsApp, WeChat, Facebook Portal, Matrix/Riot.im, iMessage, Skype is maybe back from the dead depending on who you ask (certainly not kids, though). It's almost easier to ask who isn't in the Messenger game right now. The fact that Google managed to fumble two huge market share leads between Google Talk and Hangouts is maybe only equaled by Microsoft's multiple, complicated fumbles of MSN Messenger and Skype.
But the team needs to be disciplined not to get wave overblown, that's true.
When you are a company that is as large as Google with a brand like Google, you don't want to be known to be making half baked marketing efforts, getting people on board and failing. Customers and developers then remember that and rightfully associate Google as a company that will gladly abandon you and cost you your money if it does not become big enough for Google to care (multi-billion). So Google cannot operate like a startup if they want to keep their reputation. A start-up would be very happy to have a low millions revenue but for Google that would be a rounding error and killed. So they need to make less bets, and put much more effort into getting their bets to the level that they can make acceptable money for Google while being mindful that killing a product does a lot more than lose Google the money invested.