Building a recruiting operation that scales
When every role, region, and hiring manager runs a search differently, quality depends on who happens to be hiring. I turned scattered recruiting into one way of hiring that works for any of them.
- Role
- Led recruiting, from process and systems to vendors and reporting
- Tools
- Greenhouse, LinkedIn Recruiter, Google Workspace, Slack, Trello
- Scope
- Every Orium role, across Canada, the US, Mexico, and Latin America
- Reading time
- 2 minutes
Why it mattered
Hiring gets complicated before anyone officially calls it a scaling problem. A new market needs a different sourcing approach. Another leader wants a slightly different interview. Someone builds a great scorecard and nobody else knows it exists. Meanwhile, candidates expect timely, thoughtful communication, and leaders need to know where every role stands.
When I joined Orium in 2021, the company had about 60 people. At its peak it had more than 300, across Canada, the US, Mexico, and Latin America. If hiring quality kept depending on who ran each search, growth would have exposed it quickly.
- Offer acceptance
- 95%
- Faster time-to-fill
- 25%
- More recruiter output
- 40%
- Saved renegotiating recruitment vendors
- $300K
How I approached it
My work shifted from running individual searches to building the operation behind them. The visible part of recruiting is finding people and making offers. The part that makes it repeatable is everything around that: who decides what good looks like for a role, what’s ready before sourcing begins, how interviewers judge the same things the same way, and what each leader needs to know every week.
So I made those parts consistent. Every search now starts with the same intake. Scorecards exist for each role family instead of being rebuilt every time. Interviewers assess the same written criteria. Candidates get timelines they can rely on.
- Role intake BeforeIn a DM, different every time AfterOne shared intake, the same questions for every role
- Scorecards BeforeRebuilt from scratch for each search AfterOne per role family, ready on day one
- Interviews BeforeEach interviewer asks what they think matters AfterEveryone assesses the same written criteria
- Updates BeforeHiring managers chase for status AfterA weekly update arrives before anyone asks
- Learning BeforeLessons stay in one recruiter’s head AfterThe toolkit improves after every search
And at the end of every search, a person still makes the call.
I treated reporting as part of the operation, too. Executives and hiring managers need different things, so they get different updates. Leaders see the shape of demand, where it’s at risk, and the decisions only they can make. Each hiring manager sees their own pipeline and what feedback is owed. Underneath both is a hiring demand board, updated after each staffing conversation, so it reflects what leaders said the same day.
| Practice | Open | Offers | Watch |
|---|---|---|---|
| Technology | 6 | 1 | Two senior roles past target date |
| Product | 3 | 1 | |
| Delivery | 2 | 0 | Waiting on client start date |
| Sales | 1 | 0 |
- Approve backfill for the senior developer role, or pause it to Q1.
- Confirm whether delivery demand for January is firm before we open two roles.
See the hiring-manager view and the demand board
Senior Product Manager Week 4 of search
- Two final-round panels, Tuesday and Thursday
- Offer call with the top candidate Friday
- Scorecard from Tuesday's panel, due in 24 hours
New demand 2
Sourcing 2
Interviewing 1
On hold 1
Recruiting is also vendors, contracts, and a tool stack that grows quietly until someone looks at it. I renegotiated our recruitment vendors and, after the acquisition, reviewed our systems end to end.
What changed
Offer acceptance reached 95%. Time-to-fill dropped 25% for the roles we hired again and again, because pipelines were ready before the role opened. Improving our ATS and LinkedIn integrations lifted recruiter output by 40%. Renegotiating vendors saved $300K, and consolidating systems after the acquisition saves $200K a year. Since the structured scorecards rolled out, interview feedback comes back within 24 hours instead of three to five days.
These results came from years of process, vendor, and team work. Not from any single tool, and not mostly from AI. That foundation is also what made the AI recruiting tools possible later: AI is only useful when the process underneath already makes sense.
What another leader can learn
Consistency and personality aren’t opposites. Standardize the parts that decide quality, like intake, criteria, and communication, and the human parts of hiring get more room, not less.
A process has only scaled when other people can run it without you.