Last updated: October 9, 2026
By Dave Manchester, Scientific Recruiting, KORE1
When does a growing science company stop borrowing recruiters and hire one? Hire your first internal recruiter when hiring becomes a repeating pattern instead of a run of surprises, meaning the same few roles keep reopening, two or three searches run at once, and a manager is losing days to scheduling.
That’s the short version. The longer version is that most of the founders who ask me this question are asking about money, and the money question has a better answer than the one they expect from a recruiter.
I run scientific searches through KORE1’s life sciences staffing practice, so you already know which side I’m on. Better to have that out in the open before any numbers start. What I’ll tell you anyway is that a strong outside partnership can carry a small science company for years, and then one day it shouldn’t, and the signal for that day has almost nothing to do with your headcount.
It has to do with repetition.
The Signals That Actually Mean It’s Time
Forget the employee count. I’ve watched a small analytical lab hire a recruiter two years too early, and I’ve watched a much larger contract manufacturer run for years on contract recruiters and a good talent acquisition partner without ever hiring one in-house. Headcount is the number everybody reaches for because it is the number everybody has. It just doesn’t predict anything.
Here’s what I watch instead.
| What you see | What it usually means | What I’d do about it |
|---|---|---|
| The same req keeps reopening. QC analyst, then another QC analyst, then a third. | You have a repeatable role with turnover or growth behind it. That is a pipeline, and pipelines reward somebody who owns them. | Start costing out an internal seat. This is the strongest signal on the list. |
| Three or four searches are open at the same time, every quarter. | Concurrency, not volume, is what breaks a hiring manager. Four searches at once is a job. | Internal seat, or a dedicated recruiter on retainer who works like one. |
| Your quality director is spending a day a week on scheduling and resume triage. | You are paying director money for coordinator work, and the audit prep is slipping. | A coordinator might fix this before a recruiter does. Cheaper, faster, often enough. |
| Candidates go quiet between your first and second interview. | A process problem, not a sourcing problem. An internal recruiter will inherit it, not solve it. | Fix the loop first. Hiring a recruiter into a broken loop wastes a year. |
| You just closed a round and the board handed you a hiring plan. | Money and a deadline. Usually both arrive before the process does. | Hire the recruiter. Then protect their first 90 days from the plan. |
| One role has been open eight months and nothing you try moves it. | A niche search. These are the searches internal recruiters struggle with most, for reasons that aren’t their fault. | Outside help on that one seat. Don’t build a department around a single hard role. |
Notice how many of those say something other than “hire a recruiter.” Three of six, if you’re counting.
Repetition is the thing. One hard search, however ugly it gets, is still a project with an end date on it, and projects are fine to hand to somebody outside the building. The same search four times is a function, and functions belong to people whose whole job is the function. That’s the entire argument, and everything below is just me showing my work on it.
The Four-Person Company That Earned Its Own Recruiter
Early in my career I picked up a client with four employees. Four. Not four in one department. Four people in the entire company.
My team ended up hiring across every department they built. Technical seats and non-technical ones, the lab and the office, people who would go on to run functions that didn’t exist yet when we started. For a stretch there we were the recruiting department, because there wasn’t one and there wasn’t going to be one for a while.
Then they closed a first funding round, which in a company that size changes the arithmetic of almost every decision anybody has been putting off, including the ones about hiring.
With that money they could afford something they couldn’t afford before, which was a talent acquisition team of their own. So they built one. If you’re waiting for the part where I tell you that stung, it didn’t. It was the right call for them and it had been the plan the whole time, whether or not anybody wrote the plan down.
Here’s the part that matters. They kept calling me.
Not for the roles their new team could run. For the niche ones. The searches where the whole qualified population in a commutable radius is maybe forty people and eleven of them are findable, and you need somebody who already knows which eleven. Their internal team was better than I was at hiring for their own company, because they lived there. I was better at the searches that happened once every two years, because those are the only kind I do.
That is what a partnership looks like after the internal team arrives, and it is the part of this that founders never seem to picture correctly when they are deciding whether to make the hire at all. It gets smaller. It gets more specific. It does not end.
Where an Internal Recruiter Beats Me
An internal recruiter lives in the building, and living in the building is an advantage I can’t buy. When a QC candidate asks what the night shift handoff is really like, or whether the manager who interviewed them second is the one who will actually be their boss in six months, your recruiter knows and I am reconstructing an answer from three conversations and a guess, which the better outside recruiters will admit to you if you ask them directly.
The repeatable seats are the bigger one, though. Analyst reqs that reopen every few months pay off a warm bench, meaning a list of people you spoke with last spring who weren’t ready then and might be now, and keeping that list warm is slow, unexciting work that an internal recruiter can simply do on a Tuesday instead of billing for.
Then the unglamorous half of the job, which almost nobody writes into the req. Scheduling that doesn’t take four days. Managers who’ve been taught how to interview a scientist. Scorecards somebody fills out. A hiring loop short enough that the finalist is still interested at the end of it. Your first recruiter should spend most of their first quarter in that plumbing, and if you grade them on placements over those twelve weeks you will end up firing the one person who was doing the correct work.
The benchmark data on where hires come from argues against the quiet hope behind a lot of these decisions, which is that an internal team will promote its way out of the problem. In SHRM’s 2026 recruiting benchmarking research, drawn from more than 4,600 organizations, the median share of nonexecutive seats filled by an outside hire rather than an internal promotion rose from 93 percent in 2025 to 97 percent in 2026, and the median time to fill one of those seats was 39 calendar days. Almost nobody gets promoted into these seats. Call it six weeks at the median, whoever is running the search.
What Still Won’t Fill
Then there’s the other list, and this is where small science companies get surprised.
A first internal recruiter is one person with a req load and a learning curve in your specific science. Put a genuinely scarce role in front of a person who is also covering four ordinary reqs, a benefits open enrollment, and whatever the board asked for on Monday, and the math stops working. Not because they’re bad. Because the population is small and they’ve never had a reason to map it.
Chemists are a good example of how small small is. The Bureau of Labor Statistics counted about 93,500 chemist and materials scientist jobs nationally in 2025, with roughly 6,500 openings projected per year through 2035. That is the whole country. Now filter for GMP experience, then for your dosage form, then for people within driving distance of Carlsbad or Research Triangle Park or the 128 corridor outside Boston, then for people who would actually leave. The list gets short fast. A recruiter who has never worked that list cannot conjure it in a quarter.
The seats I’d keep outside, at least at first:
- Site quality leadership, where the candidate pool is people who already have the job somewhere else and aren’t posting resumes. I wrote separately about when a small pharma company needs a director of quality, and that search is nobody’s first assignment
- Validation and tech transfer, which spike with a program and then go quiet, so an internal pipeline has nothing to feed on
- Anything microbiology. Micro benches are small, loyal, and hired through people who know them
- Confidential searches, including the one where you’re replacing somebody who still works there
- The role that’s been open eight months. Whatever that role is, it has earned a specialist
The way I work those searches is through what I call the Blind Recruit, which is not the same thing as blind hiring or anonymized resumes. Mine means the scientist I’m talking to doesn’t know which company I’m calling about. So when I ask what they want from the next five years, there’s nothing to tailor the answer to. An internal recruiter can’t do that. The company name is the first thing out of their mouth, by necessity.
What the Seat Costs, Without the Sales Pitch
The honest number is easy to look up and harder to think about.
BLS puts the median wage for human resources specialists, the category recruiters are filed under, at $75,940 as of May 2025, across about 939,700 jobs. Add benefits, a seat, an applicant tracking system, a job board budget, and the six months before somebody new is actually producing. That is your real number, and it is a fixed cost that shows up every month whether you have four reqs open or none.
An agency fee is the opposite. It is lumpy, it arrives as a single uncomfortable invoice that somebody has to approve, and it only ever appears after a person has actually started working for you. People hate it more. In a slow year it costs less.
I’m not going to re-derive the break-even here, because it’s already written up properly. Our own piece on whether recruitment agencies are worth it prices a single open seat all the way down, and the side-by-side on a staffing agency versus an in-house recruiter runs the comparison on cost and speed. Go plug your own numbers into those. Then come back with the number of hires you honestly expect next year, not the number on the board deck.
One warning about that number. Recruiter workloads have been climbing. SHRM found that the largest organizations saw median requisitions per recruiter jump 67 percent in 2026. If your plan quietly assumes one recruiter absorbs whatever you throw at them, the plan is wrong, and the person you hire will be the one who finds out.
Hire the Process Before You Hire the Person
Say you’ve decided. Good. Do these five things before the req goes up, because every one of them makes the hire work better and three of them might convince you to wait a quarter.
- Write down who decides. Not who interviews. Who decides. If that sentence takes a meeting to answer, your searches are dying of committee, and a recruiter cannot fix that from below.
- Pick the roles they own. Two or three repeatable seats, named. “All hiring” is not a job description. It’s a way to make somebody fail in public.
- Set a number for your interview loop. Days from application to decision. Any number. Having one changes behavior.
- Decide what stays outside. Write that list down too, before the recruiter starts, so it reads as strategy instead of as a judgment on them in month four.
- Agree on how you’ll judge them at twelve months. Retention, not req closure. KORE1’s 12-month retention rate on placements is 92 percent, and that is the number I’d hold any recruiting function to, internal or mine.
Then one thing about hiring the recruiter themselves, since this is a search like any other. A good recruiter with a book of relationships in your science is a person other companies want this month, not next quarter. Candidates like that behave like a house priced correctly in a good market. More than one buyer shows up, and the slow buyer loses. If your own hiring loop for the recruiter takes seven weeks, you have just demonstrated the problem you were hiring them to solve.
The Pushback I Get on This One
We’re twelve people. Is that too early?
Probably, though I’d look at next year’s plan before saying so for certain. At twelve people the better first hire is usually a part-time coordinator or an outside partner, because you rarely have enough repetition yet to keep a recruiter busy or enough process to make them effective. Ask again after the next round closes.
Could one person handle recruiting and HR together?
They can, and in small science companies they usually do, but understand what you’re trading. Recruiting is outbound and interrupt-driven. HR is compliance, benefits, and employee relations, and it fills a calendar on its own. The combined role tends to drift toward HR, because HR has deadlines with consequences attached and sourcing does not. If hiring is the priority, protect the hours for it in writing.
What does the seat actually cost us in year one?
Start from $75,940, the BLS median wage for human resources specialists as of May 2025, then add benefits, tooling, and a ramp of three to six months. The salary is the smallest surprise in that list. Our salary benchmark assistant will get you closer for your own market, and the comparison pieces above handle the break-even against fees.
Do we cut the agency loose once our recruiter starts?
Not if you want the hard roles filled. What should change is the mix. Your recruiter takes the repeatable seats and owns the process, and the outside partner narrows to niche, confidential, and spike work. That is what happened with the four-person company earlier in this post. Their new team took the roles it could run, and the niche searches kept coming to me.
How soon should we expect a hire out of them?
Expect the first placement around month three or four, and expect the first two months to look like nothing is happening. They’re building the bench and fixing your loop in that window. For context on outside speed, KORE1 averages 17 days to hire across IT searches, which works because the bench was built before you called. Nobody has a bench on day one.
Count the Pattern, Not the Heads
So, back to the question at the top. Not how many people work here. How many times has the same seat come back around, and how many searches are running right now at the same time?
If the answer is “once” and “one,” keep working with a recruiter who knows startups and spend the money on the science. If the answer is “constantly” and “four,” you’re already running a recruiting function. You just haven’t hired anybody to do it yet. And if you’re somewhere in the middle, which is most companies, the honest move is a hybrid: somebody internal who owns the roles you repeat, through direct hire or a contract-to-hire trial, with the rare searches staying outside where the relationships already exist.
Trying to work out which one you are? Let’s Connect. Tell me how many seats you filled last year, how many came back open, and what’s been sitting unfilled the longest. I’ll tell you whether you need a recruiter, a coordinator, or a better interview loop. Sometimes the answer costs you nothing. You can also find me on LinkedIn.
