There's so much this article misses: - most large companies a/b test their redesigns. If your company isn't big enough (for the tools or traffic needed to a/b t

There's so much this article misses: - most large companies a/b test their redesigns. If your company isn't big enough (for the tools or traffic needed to a/b test), then yeah, you're left to judgment calls. Really good a/b tests will actually capture things like "existing users initially hate it but eventually end up using the product more" - a lot of UI redesigns are driven by effect on new users, which might be more important to the company. Think about it, if you're already using the product, you probably don't care what it looks like too much (and might be annoyed if it changes). For new users, anesthetics leave a lasting first impression. Yes, at many small companies without data, redesigns can be a waste of time and even harmful. But that doesn't mean they should always be avoided.