They've killed APIs/Products, but they've almost always given a very long amount of time to transition.

Also have killed APIs/Products that are extremely difficult to transition from with short notice. For example, they gave ~4 months for the Google Wallet for the web transition, and no data portability for businesses with users on subscriptions.

Stripe, for example, has a data portability clause that allows you to move card data to another payments processor that meets some compliance standards.

https://support.google.com/payments/merchant/answer/6107573?...

Regardless of if you get 4 months notice or 1 year notice you still have to do the work to migrate and you're still stuck with the sunk cost of investing in/learning a doomed platform.

That's why I wouldn't build a business on Google, because they have a long history of killing things when they aren't wildly profitable/successful. A $10m/yr profit product is considered a distraction of valuable engineer time unless it has some ulterior goal for the company.

And I'm not saying they should change -- they do what's right for them. One engineer working on some distraction project could instead be moved to ads quality and end up making a change worth anywhere from hundreds of millions/yr to billions.

But they can't have their cake and eat it.

Google App Engine comes to mind. I'm amazed they would kill of such a product at all.

Google App Engine isn't killed. It's still alive and new features are being added. What are you talking about?

The GP was probably referring to the huge price change that effectively killed it for many people.