Days 1 to 30: the hiring process, built before the next req opens
This phase exists because the most expensive People mistake at this stage is hiring fast without a repeatable process, which produces inconsistent bars, slow decisions, and candidates who ghost because the loop dragged on too long.
Build a scorecard for every open role defining what success looks like in the first six and twelve months, not just a list of skills, and set a rough comp band for each role at the same time. A wide band anchored to two market reference points is enough for now. The point is that the first offer you make under the new process already gets checked against something, since refining bands is a phase-two job but having one to check against is a phase-one job. Design the interview loop, three to four stages maximum for most roles at this stage, each stage testing something distinct rather than everyone re-asking the same questions. Write the offer process, who approves what, how fast an offer goes out after a final yes (speed matters more at this stage than most founders expect, because strong candidates have other active processes). Set up basic onboarding, a first-week plan so a new hire isn't improvising their own ramp-up.
Days 31 to 60: comp structure and documentation
By day 30 you have rough bands and may have made an offer against them. This phase turns those rough bands into a real structure and builds the documentation layer underneath. It matters here because this is where ad hoc decisions from the first few hires start compounding into inconsistency that gets expensive to unwind.
Turn the rough bands into real ones for every role and level you currently have, using the three-band structure in Comp Banding and Leveling for a 10-35 Person Team. For roles you haven't hired yet, set the band from market reference points and expect to revise it once you've run a search and seen what candidates actually ask for. Write an employee handbook covering the policies that actually exist in practice, not a generic template, PTO, remote work expectations, code of conduct. Set up an offboarding checklist, this matters earlier than most founders think, because the first departure (voluntary or not) at a company this size is usually also the first time anyone realizes there's no process for it.
Days 61 to 90: the systems that scale past the first ten hires
The last phase is what turns ad hoc decisions into an actual function that survives the founder no longer being personally involved in every people decision.
A lightweight performance check-in cadence, doesn't need a formal review cycle yet, but regular, structured check-ins beat "we'll talk if there's a problem." A basic People metrics view, headcount plan versus actual, time-to-fill per role, early attrition if any, enough to answer a board question without scrambling. A plan for what's next, usually either a part-time or fractional People function continuing to run this, or a first full-time People hire, decided deliberately rather than by default once the founder runs out of bandwidth.
What this deliberately doesn't include yet
A full performance management system, a formal leveling framework beyond basic bands, or a dedicated People Ops tool stack. All three are real needs eventually, but building them before you have the hiring process, comp structure, and basic documentation in place is solving a problem you don't have yet while leaving the one you do have unsolved.