And products just languish as a result.

The theory isn’t anything deep, but it seems to fit the facts.

The data used by the person in charge of making this decision (typically a VP or a Sr Director): what is the probability of them getting promoted. If doubling down or just maintaining and polishing existing projects was rewarded, Google wouldn't have been in this state.

Now why those incentives exist that way? It was a confluence of factors in 2000's when internet was growing rapidly, Google had lots of talented employees, a bold bet in a new area (email, news, street view or browser) had a higher expected value than polishing and maintaining existing products. The culture needs a reset, Google is currently living through its "Ballmer era".

polish/maintain -> stagnate

You can manipulate data to show whatever you want. The incentive structures within the company are to blame.

Edit: Limp leadership is also to blame. Probably some other things too?

A lot of times it's much easier to find data to give you reasons not to do something long-term (very similar to investors focused on short-term results) than to confirm the value of committing significant resources to long-term bets.

Other than "keep milking ads" is there any clear strategy Google's consumer business has shown? Android looks like a good example of executing on that ads strategy, but other things that seem harder to tie to ads directly tend to languish since nobody there actually knows what they want to do in other areas: chat apps? Stadia? Random things like Cardboard? Consumer G-suite? I'd say even assistant/home stuff is languishing (where's the big money maker).

They're not a data driven company, they are a promotion driven company. New products drive promotions and maintaining existing products do not.

Could be as simple as “We tried this for x months. It still isn’t profitable, and we don’t see it becoming profitable soon”

For me, with zero inside knowledge, it looks like Google’s upper management gives managers lots of freedom to launch products, but expects them to bring in money soon.

That doesn’t seem the case for all kinds of products, though. I don’t see Chrome directly making money for them, for example. There, it seems they realize it indirectly, by defending their advertising income, is profitable.

It seems upper management doesn’t have a long term strategy for making money with messaging apps, but still has hope some of their managers will find it.

When the data shows the results you are looking for, you say you're data-driven. When it doesn't, you manipulate and torture it long enough to produce the result you are looking for. If it still doesn't, you come up with some excuse to end up doing the thing you wanted to do anyway.