There's an illusion you can slowly roll out major changes and somehow people will adapt. If you offer a way to slow down or keep the old version around "just as an option" (everyone loves options!) then you're supporting the old thing forever. Why? 1/

2/ When does the adjustment period end? If people are paying you there is one answer: NEVER! Plus if you offer an adjustment, big customers will just tell you "Great, we will let you know when we're ready to adjust." If there is any negative feedback that will be: NEVER!

3/ The illusion of "options" will only work one time. Why? Because by the time you want to make the second change some people will still insist on the old way. Then you have to figure out how the new new features fit in the old experience.

4/ This dynamic is why for the past decade we've seen so few major overhauls of SaaS products. Certainly in consumer you can see the uproar over even the smallest changes. No uproar? That just means no one cares enough or uses it enough. But but you say..

5/ Small changes are made all the time? YES! But the problem is x number of small changes over time is not a coherent design and at some point the whole system is just a big mess.

6/ Apple is experiencing this right now with Liquid Glass. The calls for "a slider" or "option for flat" or whatever. The net of all this is that when you make a change you have to commit. If you don't commit and think "oh an option will be fine" you're basically uncommitted.

7/ Working against you the whole time are the forces of engaged and vocal tech people who ALWAYS LOVE OPTIONS. And they all say options are low cost. Options are easy. The old code is there. Just give US, the PEOPLE, CUSTOMERS, the choice. It's brutal.

8/ Yes lived this with Office and Windows 8. What we did in Office is why we took the approach with Windows 8. Be critical of that or not but there's not a path down the middle. No matter how much people say there is.

9/ PS: This applies to API changes and UI changes. API changes are even more difficult. This is something Apple has done extraordinarily well. But at current scale they are seeing how tricky it can be. x86 probably stuck around longer than they expected. Microsofts approach was always to keep everything forever. That has costs too.

10/ More on Office here:

11/ More on Windows here:

PS/ The bottom line for all this is in the early days there are tech enthusiasts. They love change. They embrace it. Then suddenly they don't want change right now because a product is important to them. Products in general might be central to people but not central to their lives. They always have priorities other than being excited about the latest thing to change in some tool they use, no matter how important. You have to respect that.

PPS/ Want to read a Harvard Business School case on this, here's one written by my good friend @marcoiansiti about Microsoft Office. ($)

@JoannaStern PS. I will die on the “let’s stop anthropomorphizing” hill.