The Citywide Rewrite I Talked Myself Out Of Building

Casa MD is a Chicago health services app for people who've lost coverage and need affordable care fast. A newer draft of the plan proposed going citywide: 195 sites, a rebuilt organizations table, new city data sync, a rewritten routing algorithm. Well-reasoned. Not what the current build actually needed.

The real problem was narrower than any of that. Someone who'd just lost free coverage was being told to pay $35-55 out of pocket at a clinic, which just restates the loss instead of solving it. A real user, or a judge watching a demo, would catch that gap in about five seconds.

So I scoped it down to the actual fix: surface a real coverage program as the answer to that specific loss, seeded with verified data, shown ahead of the paid clinic list. A few hours of work instead of a full data-layer migration, a new organizations table, and a rewritten routing algorithm that a citywide rollout would have required.

The bigger, citywide version might genuinely be worth building eventually. Casa MD covering 195 sites with a real routing algorithm behind it is not a bad idea. It's just not this week's idea, and definitely not something to stack on top of an unmerged mobile pivot, days before a demo, when the smaller fix would have caught the exact same gap a real user or a judge would notice first.

That's a different kind of scope discipline than resisting a shinier, unrelated idea, which is its own well-known trap. This was about resisting an over-engineered version of the same idea, the one that looks like more thorough problem-solving but is actually a way of avoiding the smaller, less impressive fix that's the one actually needed right now. The citywide version feels like doing the job properly. The narrow fix is what actually stops someone from being told to pay $35-55 they don't have.

If you're staring at two versions of a fix, the narrow one that solves what's actually broken and the sweeping one that would be more impressive to build, ask which one a real user would notice first. Build that one. Write the bigger version down for later instead of letting it eat the days you don't have.