I remember hearing a talk by Bill Gates a long time ago where he said the biggest competitor for Word was previous versions of the program. That is no longer true with forced updates. Granted, there are many reasons why it is not practical these days for old versions to be a competitor, but that would be a good message to companies that some of their new "innovations" may not have been a good idea.
I think this is one of the biggest issues with SaaS. I appreciate that a stable revenue source is necessary for ongoing development, but UX decisions are much harder to evaluate. If I purchase a desktop application and opt not to upgrade, that's useful data to have. If you jam a UX update down my throat and I don't cancel my subscription, that doesn't mean you made a fantastic UX decision. It generally means the cost of changing is too expensive or, more likely, you've locked my data up in a format not easily exportable.
I think it'd be nice if our tooling made it easier to support multiple "versions" of a deployed application. As it stands, it's far too expensive and error-prone to let customers opt into UI upgrades. As a product consumer, I can't stand continuous deployment. It just means my workflow can break at any time.
Competition forces you to make a good product or die and be replaced by a company that does. When previous versions were a source of competition, it was likely easier for companies to notice their mistakes and correct them. With SaaS, a company can coast along on brand recognition for awhile, slowly become zombified, until one day all of your customers switch to an up-and-coming competitor.
