Every June I usually go out back and cut the suckers off my tomato plants, those little shoots that pop up where a branch meets the main stem. This year I didn’t. June got busy, July got busier, and by the time I actually walked back there in early August, the bed didn’t look like a garden anymore. Leaves piled three deep, vines swallowing the stakes. Half the tomatoes I only found by accident, already gone soft because nobody, meaning me, ever saw them ripen.
Not the tidy version I’d usually pick to show off. The one where I clearly dropped the ball.
Some of why is obvious enough. My son is approaching two this summer, and my Saturdays stopped being mine right around then, not for anything with a name like a pruning schedule. Most of them go to him hunting through the vines with his toy scoop like it’s a metal detector at the beach. Credit where it’s due, that might be a better use of a Saturday than the pruning ever was.
Standing in the middle of that mess anyway, equal parts annoyed at myself and charmed by him, I got to thinking about why I let it happen. I didn’t decide to skip pruning. I just skipped one Saturday, then skipped the next one too, and by the third skipped Saturday, skipping was just what Saturdays were. Which is exactly how a product ends up buried under its own growth too. Nobody sits down and decides to stop cutting. They just skip the ugly conversation once, then again, and skipping becomes the whole strategy.
I’ve never once regretted cutting a sucker. I regret almost every feature I didn’t cut. Turns out you can also regret the ones you just plain forgot.
I’ve watched this play out more times than I can count, from both sides of the table. Roadmap’s up, or the app is open, and somebody floats cutting a feature nobody uses and nobody wants to touch. The Zoom goes quiet. Not thinking quiet. The quiet of somebody bringing up politics in front of the grandfather who’s been waiting all year for the opening (too specific?). Then somebody says it:
“Well, we already built it.”
I used to write that hesitation off as laziness, a simple refusal to do the work. But the truth is more complicated. We hold onto features because we have a hard time letting go of the investment they represent. Building used to cost real time, a whole sprint, somebody’s actual Tuesday. Once you had paid that price, keeping the thing felt like the only responsible choice. Now a model can spit out a feature in the time it takes to make a pot of coffee, but our instincts have not caught up to that reality. We are still clinging to the weight of costs that we no longer actually pay.
I watch this happen constantly in AI-built products now. I watch it happen to us too, if I’m honest. Matt calls it the addition reflex: when generating’s nearly free, every problem starts looking like it needs one more thing bolted on, never something cut. He’s said it plainer than I ever have in Why AI Makes It Worse. This one’s about the other half, the part where you’re standing in your own backyard wondering how you let it get this bad. Nobody adds a bad idea on purpose. They add a reasonable one, forty separate times, and the product ends up looking like it was assembled by forty people who never sat in the same room. Which, again, is often exactly what happened, just with the seat count doubling as the prompt count.
You can add a feature and never once decide what your product is even for. It just sits there, defensible on its own, agreeing with nothing. Cutting is different. The moment you pull something out, you have to say why, out loud, in front of the room. This is what we’re building, this is who it’s for. Everything else was somebody else’s product wearing our logo. That’s the freeze point for most teams, and it’s rarely laziness or ego behind it. Cutting drags that conversation into the room whether you like it or not. Skipping it and calling it momentum? Way easier.
You know the feature I mean. You knew it before I finished that sentence. The one that shipped because a competitor had it. A hostage situation with better lighting, not a reason. The settings page nobody can explain. The file from week one everybody’s afraid to touch. The real reason you’re keeping any of it is that it’s there. Not because it’s good. Somewhere along the way we all agreed that being there was argument enough. It isn’t.
When we audit a product for a client, most of the work is just triage. Keep, consolidate, delete. Five button styles become one, so four of them die on the spot. Three tones of voice become one, and somebody’s favorite line of copy goes with it. What’s left usually looks smaller on the feature list, and reads like one person made it on purpose instead of a group project nobody proofread.
I fall for this too, more than I’d like to admit. I’ve kept screens alive out of something that felt like loyalty and was just guilt about the hours somebody put into them. The tell is the same one I use on clients now: if the only reason you’re defending something is that it exists, it’s a sucker. Cut it.
I’ve never once been sorry I cut a sucker. Every feature I didn’t cut, I’m sorry about. I have to remind myself of that one every single June, apparently, including this one.
Your users never see the cutting, same as nobody’s going to see this mess once I actually get back out there this weekend. They’ll just see a garden. Most products already have plenty of good ideas lying around. Somebody willing to actually walk the rows is the harder thing to find.