COO, VP Operations, Head of Operations, or General Manager. Available now.
Remote or Seattle · Full-time or contract
I'm most useful at multi-location field-service businesses and home-services portfolio companies — the operations I know from the inside.
Why is an owner-operator looking for a role?
Because it doesn't need me day to day anymore, and I'd rather work on a harder problem than protect an easy one.
Thirteen years in, NW Maids runs on systems and a three-person back office. I built it that way deliberately, and it worked — which means my actual job now is a few hours a week of judgment calls and approvals. That's a good outcome for the business and an underuse of what I know how to do.
I'm not leaving. I still own it and it still runs. But the interesting version of this work is doing it somewhere with more sites, more people, and more at stake than a single five-city operation — and that's a job, not a side project.
P&L ownership — thirteen years, through growth to roughly $5M annually.
Multi-market expansion — five metro markets, each with its own staffing pool, demand pattern, and local search footprint.
Remote team building — a three-person back office running scheduling, support, hiring, and field-issue resolution across all five markets. No physical office.
Field workforce at scale — 100+ cleaners, with the recruiting, onboarding, scheduling, and quality problems that come with a distributed hourly workforce.
Operations infrastructure — the two builds documented on this site: a full SaaS-to-self-hosted migration, and an AI operations layer with enforced human approval.
I'd rather lead with the second part.
Transfers cleanly: the written-constitution pattern for governing any automated system. The staged-approval model — analyze automatically, execute manually. The reporting layer: lead attribution, rank monitoring, ad-lead reconciliation, no-reply detection. Deterministic integration plumbing. Hosting economics that don't scale linearly with site count.
Does not transfer: my approval queue is a single file that one person reads — a bottleneck at twelve locations. Policy-based guardrails work with one owner-operator and break with multiple. Several integrations run on logged-in browser sessions, which is fragile and not auditable. And there's no queryable audit trail by action class, which you'd need for insurance rather than curiosity.
The full version of that analysis is in the second case study.
Days 1–30. Learn the actual operation, not the org chart. Ride along with the field. Sit in the support queue. Find out where the back office is spending time that nobody has looked at in a year.
Days 31–60. Baseline everything before changing anything. This is the mistake I made in my own business: I automated first and lost the ability to prove what changed. Capture the numbers while they're still ugly.
Days 61–90. Fix the highest-cost manual process, with a documented rollback. One visible win the team trusts beats five they didn't ask for.