[DDayMace]

There's a lot of truth to this. Back when I was doing UI for a big bank, I urged them to cleanup their frontend technical debt and unify their interface into a more consistent modern look and feel. They ignored it and I ended up leaving rather than hack into jQuery all day.

They understandably emphasized stability over replacing tested components, which financial is often known for. However I think their visible sign of improvement to the customer is severely lacking even today (this was five years ago), especially compared to their competitors. My overall feel is that the site is brittle regardless of what's happening under the hood.

[rendaw]

At some point you're making a choice between old users and new users. And the number of old users will only decrease if you don't have new users.

[gedy]

Strategy I've taken in some places is: do not loop in UX or advertise this is a big rewrite. UX, PMs, and some engineers just cannot resist the opportunity to change the UI around and your important tech debt works gets blamed for the rest of the org's fiddling.

[DDayMace]

Yes, that's a good strategy. Ask for permission rather than forgiveness. A former colleague and I did just that for awhile, and we were able to avoid the politics and fit the work into sprints.

[ldjkfkdsjnv]

Large established companies cannot meaningfully change their UI without alienating all of their legacy customers.

[DDayMace]

It seems like if the company did a bunch of A/B testing on a wide range of customers, they could figure out if such legacy customers were truly satisfied with the existing product, or perhaps prefer the new UI.