By the start of this year I had eleven plugins installed in my design tool, each one solving a real problem I’d had at some point. I could name the exact frustration that justified each install. What I couldn’t do, if you’d asked me cold, was name more than four I’d actually used in the past month.

The cost that doesn’t show up anywhere

Every plugin is small on its own. An extra menu item, a slightly longer load time, one more update prompt to dismiss. None of that shows up as a cost in any way you’d notice day to day — until you try to explain your setup to someone else, or open the tool on a machine that doesn’t have your accumulated eleven plugins, and realize how much of your speed was quietly borrowed from tools you’d forgotten you were relying on.

The actual cost was decision fatigue I’d stopped noticing because it had become ambient. Eleven plugins means eleven places a keyboard shortcut might conflict, eleven extra icons competing for attention in a toolbar, eleven small “which tool handles this” decisions per day that I’d stopped consciously making but hadn’t stopped paying for.

What happened when I turned them off

I disabled everything except the four I could name using in the past month, expecting to immediately need three of the seven back. I needed one, once, for a specific export task, and reinstalled it for the afternoon. The other six I have not missed, which is the part I found genuinely surprising — I’d have bet real money at least half of them were load-bearing.

What I actually got back wasn’t speed in any single task. It was a smaller, calmer surface — fewer places to look before doing something routine, fewer decisions about which tool handles a job four tools claim to handle. The gain wasn’t in any one workflow. It was in the space between workflows, which had gotten quietly cluttered without a single moment where that clutter felt like a decision I was making.

The actual lesson

I don’t think plugins are bad, and I’ve reinstalled a couple since, deliberately, for specific recurring needs. What changed is the default. I used to install first and ask later whether I still needed something. Now the default is off, and something has to earn its way back in by solving a problem I’m actually hitting this month — not a problem I hit once, eight months ago, that justified an install I never revisited.

Rhys O’Sullivan is a product designer at Fieldstone, previously Loop. This is one in an occasional series of notes on research, systems, and the process of making things.