> unsanctioned side projects
Wrong focus. Unplanned, but valuable and necessary side projects, that focus on the long term, setting aside for a moment that permanent focus on the short term next milestone next sprint next story next task next function next line of code cyclic trap of short term focus - that product managers (along with literally everyone else, myself included) fall all too easily into.
If my employer doesn't trust me to have some sense of what constitutes valuable and necessary, why the hell did they hire me in the first place? And why do I want to work there? And how soon will I be laid off, fired, or quit of my own accord - leaving for greener pastures where there's higher morale, and less micromanagement? These things will translate directly into better productivity, in turn translating into moving up the career ladder faster?
(I should note: I pick jobs where I consider my "20%" projects to be business relevant. That doesn't mean they're sanctioned.)
> I'd take some of my spare time
1) I can't focus on proper work projects outside of work. I've tried it. Even if I could, reducing my weekends to 1 day a week sounds like a great way to burn out. If somehow I don't burn out, bluring the line between work and play enough for the last day of the weekend I kept for myself to never quite feel like I've left work at work.
2) Prototyping is insufficient.
3) What spare time?
> writing code that will be owned by your employer
That's a feature. Do you have a copyright assignment clause in your contract for all your home projects? Is it legally enforceable in your jurisdiction? Nebulous. Here's the deal: I work on things at work, they get to own it - clearly and unambiguously. Best case scenario? I get recognition, praise, raises, bonuses, promotions, and they ask me to work on it even more. Worst case scenario? I'm off on an adventure, in search of employers that better value my talents. Win/win situation.
> all the while taking longer to do your assigned tasks and looking like you slack off for a day a week?
If it looks like you're slacking off, you're doing 20% time wrong. I have no games open, no social media, no reddit, no HN - my 20% time is sit down and write some code time. You'll hear my keyboard, and you'll see code if you look at my screen. In the lulls, you'll see me thinking hard, unaware of the world around me.
Even if I made it as regular as one specific day a week (I don't - it's hours here and there, fit in and around the normal ebb and flow of my daily work) I'd wager that fluctuation in stats-measured productivity would still get drowned out by the standard deviation of noise in normal work. Some days I commit 10, 20 things with relative ease - other times a single obscure bug wastes 2 weeks of my life to track down - and you're worried about a day a week?
> If they're going to be happy with the project then it's better to do it without pissing off everyone else.
Here's the thing: 20% time (possibly rogue, possibly unsanctioned) is how you pull this off. Which is easier to plan project schedules around: Consistent 80% time spent per week on the project, or sometimes 100% and sometimes 0%? Which sells better:
A) "I want to add an unplanned week to the project schedule so we can replace some terrible tools that we've been limping around with. No, it's not a customer requirement, why do you ask?"
Or:
B) "By the way, I wrote a spiffy new tool that replaces that old terrible one. We can now get those customer requirements that require the tool done faster. You're welcome!" (Unmentioned: You implemented it in about a week, amortized over the previous month or so.)