Your first 90 days in a tech ops role: a runbook
Starting a new ops role is a strange mix of drinking from a fire hose and feeling like you have nothing to do yet. The temptation is to prove yourself fast by fixing everything you can see is broken.
Don't. Not yet. The best first 90 days I've seen follow the same shape as a good incident response: observe before you act, find the real root cause, then fix one thing properly.
Days 1–30: listen and learn the system
- Meet everyone who touches your work. Your team, the teams upstream and downstream, the frontline folks. Ask each of them: "What's the one thing that makes your job harder than it needs to be?"
- Read the docs your team actually works from. Which ones depends on the role:
- Support ops and escalations: macros, help center articles, and the escalation path.
- Incident management: runbooks, the on-call guide, and the last few postmortems.
- Trust and safety: the policies and the enforcement guidelines behind them.
- Program management: the roadmap, project plans, and the last few status updates.
- Vendor management: contracts, SLAs, and vendor scorecards.
- Community ops: community guidelines and the communications calendar.
- Learn the numbers. What does the team measure, how is it calculated, and where does the data live? Get access to the raw exports, not just the dashboards.
- Shadow the work. Sit with the people doing it. You'll learn more in two hours of listening in than in two weeks of docs.
> rule for month one: write down everything that seems broken. Fix none of it. Some of it is broken for a reason you don't know yet.
Days 31–60: map it and pick your first fix
- Draw the process end to end. Every handoff, every tool, every place a ticket can get stuck. Share it and let people correct you. That's how you find out what's really happening.
- Look at your month-one list again. Which problems came up in more than one conversation? Those are your candidates.
- Pick one early win. Small enough to finish in a few weeks, visible enough that people feel it. A broken macro, a confusing escalation path, a report someone rebuilds by hand every week.
- Agree on it with your manager before you start, so it counts.
Days 61–90: ship it and write it down
- Ship the fix and tell people what changed, why, and what to do differently.
- Measure the before and after. Even a rough number beats "it feels better."
- Document it. Update the runbook so the fix outlives you being the person who knows about it.
- Write a 90-day recap for your manager: what you learned, what you fixed, and the three things you'd tackle next. It becomes your plan for the next quarter and a head start on your review.
The mistakes to skip
- Saying "at my last company we..." more than once a week.
- Fixing the loudest problem instead of the most common one.
- Building relationships only upward. The frontline team will tell you the truth your dashboards won't.
If you're still in the interview stage for your next ops role, my free interview prep kit has a story bank and the questions these roles actually ask. And once you're in, the KPI tracker is a good home for those before-and-after numbers.
Starting somewhere new soon? Tell me the role. I love this stuff.