Cleaning Insights

The Thesis

This one started with a good friend, not a data point. He manages a portfolio of short-term rentals and a roster of cleaners across them, and he was living a problem I found genuinely interesting: how do you actually manage a distributed team of cleaners across rental turnovers — vetting quality, scheduling around guest checkouts, handling payment, and reacting fast when a same-day emergency threatens the next check-in. That’s an operations and trust problem hiding inside an unglamorous, high-frequency job.

What I Built

Rather than start from a feature list, I started from his actual week. We talked through how he currently hires and evaluates cleaners, how he checks quality after a clean, how scheduling breaks around last-minute guest changes, and where the real emergencies come from. The idea was a SaaS layer for cleaning companies and property managers that turned that whole workflow — job assignment, post-clean verification, payment, emergency escalation — into something structured.

What Happened

The pain was real, but what the research surfaced was a gap between “this is genuinely annoying” and “there’s a software business here” — one operator’s frustration, however real, is a sample size of one, and we hadn’t found evidence yet that the willingness to pay for software (versus just hiring better, or managing it by hand) existed beyond his situation. I kept it in research rather than pushing it into a build because that gap never closed.

What I Learned

One operator’s real pain is a sample size of one — check for a repeatable, paying market before writing any code.

• A friend’s vivid frustration is a great place to start research, and a risky place to stop it.
• “This is annoying” and “people will pay for software to fix it” are two different claims — test them separately.
• Staying in research isn’t a lesser outcome than shipping — it’s the same discipline, applied honestly.