If your new product with ten years of improvements was so unattractive compared with the existing version your customers are currently using that they were choosing not to update, what does that say about the value to the customer of all those improvements?
In the world before SAAS we saw both extremes often. Some products had incredible dev teams and people would upgrade relatively often even when each new version cost extra money. Other products barely produced an interesting change in several new versions over several years and users mostly didn't pay for those updates.
SAAS is a convenient business model for the developers in the latter category because it locks in the ongoing revenues whether any future changes add value or not. They just need enough lock-in effects that once you've signed up you can't easily move to one of their competitors.
This is one of the reasons business productivity software becomes unwieldy after a while -- it's fine to add buttons and menus, you just can't change existing ones or what they do. So then version 5 will finally introduce the massive, desperately needed UI refactor after ten years, while plenty of clients just stay on version 4. (An example of this is the Blackbird CMS -- old Blackbird vs new Blackbird.)