Should You Use an ATS or a Recruiting Service?
Software, agency, or a platform with an embedded recruiter: the decision turns on whether anyone owns hiring as their job. When each model works, and what it costs.
Buy software if someone in your company already owns hiring as part of their actual job. Buy the outcome from an agency if nobody does, you need two roles filled this quarter, and you can live with owning nothing when it is over. Take the hybrid, a platform plus an embedded recruiter, if you need both and cannot yet justify a full-time hire. The decision variable is not budget. It is whether a named person wakes up on Monday responsible for the pipeline.
I ran a recruitment agency for four years before I built Pickr, so I have sold the service and I now sell the software. That gives me two conflicts of interest rather than none. What follows is what I would tell a founder over coffee, including the cases where I would tell them to sign with an agency and stop reading.
Why an ATS with nobody to run it becomes a database
This is the failure mode I see most often. A company buys an applicant tracking system in January because a board member asked how the hiring process is structured. By March it holds a few dozen candidates, some of them duplicates, and nothing has moved past the first stage in 6 weeks. Nobody lied and nobody was lazy. The CTO who bought it had a product to ship.
An ATS does not source. It does not chase a hiring manager who has not opened a scorecard. It does not write the follow-up to the candidate who went quiet. It removes friction from work that somebody still has to do. Cut a process that consumes 8 hours a week down to 3 hours and you have saved 5 hours. If nobody had 8 hours to begin with, you have saved nothing.
The honest test: name the person. If the answer is "we all pitch in", or "the founders" while the founders are also selling and building, you are buying a filing cabinet.
When running hiring in-house on software actually works
It works more often than agencies would like you to believe. Three conditions:
Someone owns it. A founder who takes hiring seriously for a 6 to 10 month stretch, an office manager who grows into a coordinator role, an engineering lead who genuinely enjoys interviewing. One day a week, consistently, beats a heroic push every second month.
Your roles are findable. If people apply to you, or your network reaches the profile, sourcing is not the bottleneck. Structure is, and structure is what software fixes.
Your volume is low but steady. 4 to 12 hires a year is the sweet spot: enough repetition to build a pool and learn what a good hire looks like, not enough to drown one person.
Below roughly 4 hires a year, a spreadsheet and a shared inbox will not kill you, though you will have no memory of who you rejected. The talent pool you build over time is the compounding asset, and it only compounds if it sits in one place.
What a recruitment agency buys you, and what it keeps
An agency buys you throughput and a market map, immediately. A good recruiter has spoken to dozens of people in your niche in the last 3 months and knows roughly what they are paid and who is restless. You cannot buy that with software, and a vendor who tells you otherwise is selling.
Pricing in Germany and Austria is fairly consistent: 20% to 25% of first-year salary on contingency, so a 70,000 euro role costs about 15,000 euros. Retained search costs more and tends to behave better, because the recruiter is paid to work the mandate rather than to win a race against two other firms. On a live, well-briefed search, 30 to 45 days to a signed offer is a reasonable expectation.
The cost people underestimate arrives at the end. When the contract closes, the interesting part stays with the agency: the dozens of people who were nearly right, the reason the third candidate declined, the salary expectations of everyone who said no. You have the hire. You do not have the search. Run the same role again 2 years later and you start from zero, and you pay the fee again.
That is not agency malice, it is structure. Their database is their asset, and handing it over would be handing over the business. If it matters to you, negotiate an export clause before you sign, and keep your own candidate pipeline running in parallel so agency-sourced candidates, referrals and inbound applicants land in one place under one evaluation standard.
What recruiting as a service is, and what you give up
The third model is a platform plus an embedded recruiter who works inside your process rather than selling you finished candidates. That is what recruiting as a service means when it is done properly: your system, your pipeline, your data, someone else's hours and craft.
What you give up is worth naming. You give up some control over sequencing and priority, because the person is not a direct report and is not in your Monday meeting. You give up the belief that your process is unique, since an embedded recruiter will push you towards a structure that works across companies rather than the one you invented. And on the hardest specialised searches, a CFO, or a first sales leader in a market you do not know, a retained specialist with 15 years in one vertical will still beat a generalist embedded recruiter. I would not pretend otherwise.
What you keep is what the agency model takes away. The nearly-right candidates stay yours. So does the reasoning behind every decision.
ATS, agency or embedded recruiter: which fits your stage
| Stage | Hires per year | What I would do | Why |
|---|---|---|---|
| Under 15 people | 1 to 4 | Founder-led hiring on a light ATS | Volume is too low for anything else; you mainly need to stop losing people in your inbox |
| 15 to 50 people | 5 to 15 | Platform plus embedded recruiter | Too much for a founder, not enough for a full-time hire; you need the pool for later |
| 50 to 200 people | 15 to 40 | In-house recruiter plus platform, agency for the two hardest roles | A recruiter pays for themselves against fees above roughly 15 to 20 hires |
| 200+ people | 40+ | In-house team, retained search for executive roles only | Scale justifies the function; buy specialists, not throughput |
| Recruitment agency | Any | Software, always | Hiring is the product; outsourcing it makes no sense |
The row people argue with is 15 to 50 people. The instinct there is to hire a recruiter early because it feels like ownership. If your hiring is lumpy, 8 roles in one quarter and none in the next two, a full-time recruiter spends half the year underemployed and starts building processes nobody asked for. I have written elsewhere about running in-house hiring without headhunters; the short version is that it works when volume is steady, not when it is spiky.
Where Pickr fits in the software-versus-service decision
Pickr is an AI-native recruiting platform for recruitment agencies and in-house teams. It scores every candidate on evidence of skills rather than keyword matches, including adjacent and transferable skills, and what happened to the people you actually hired feeds back into how the next candidates are evaluated. Interviews are transcribed and scorecards arrive pre-filled with evidence mapped to each criterion, so an interviewer edits a draft instead of facing an empty form days later. Interviewer and hiring-manager seats are free, which matters because the people you most need in the loop are usually the ones least willing to log into anything.
For agencies, multi-client and multi-brand pipelines, client portals and placement tracking are native rather than permission workarounds. Candidate data sits in Frankfurt, the product is built in Austria, a DPA is included, and personal data is redacted from AI prompts by default.
Pickr can connect to your current ATS read-only for an audit, which is a low-commitment way to find out whether your problem is really the software, and it can import your hiring history if you move across. We do not publish list prices, because what a 3-person agency needs and what a 50-person talent team needs are not the same conversation.
What Pickr does not do is decide who owns hiring. If nobody in the company is accountable for the pipeline, better evaluation produces a better-organised stall, and you should sign with an agency instead.
How to make the call
Do not decide between software and service on price. Decide on ownership. If someone owns hiring, buy the platform and give them the tools to compound what they learn. If nobody owns it and the roles are urgent, buy the agency outcome, but write the export clause first. If you sit in the middle, where most companies between 15 and 50 people sit, the hybrid is not a compromise: you rent the hours and keep the asset.
The wrong move is buying software as a substitute for deciding who is accountable. That decision costs nothing and changes more than anything on the shortlist.
Frequently Asked Questions
Should a startup buy an ATS or use a recruitment agency?
Buy software if someone in the company already owns hiring as part of their job and can give it roughly a day a week. Use an agency if nobody does and you need roles filled this quarter. Software with nobody to run it becomes a database of stale CVs within 2 months, and the fee you saved comes back as roles that never close.
What does a recruitment agency cost compared to recruiting software?
Contingency search in Germany and Austria usually runs 20% to 25% of first-year salary, so one hire on a 70,000 euro salary costs roughly 15,000 euros. Recruiting software costs a fraction of that per year, but it only pays back if a named person is actually using it every week. Unused, it is more expensive than the agency, because the roles stay open.
What happens to my candidate data when an agency contract ends?
Unless you insisted otherwise in the contract, usually nothing comes back to you. The interview notes, the rejected-but-good candidates and the reasons behind each decision stay in the agency's system, because that database is the agency's asset. Ask for an export clause before you sign, and run the pipeline in your own system in parallel if the relationship matters.
How many hires a year justify an in-house recruiter?
Roughly 15 to 20 hires a year is where a dedicated recruiter starts to pay for themselves against agency fees in Germany and Austria. Below that, either a founder who genuinely owns hiring plus good software, or an embedded recruiter on a monthly retainer, is usually cheaper and faster. Lumpy volume matters as much as the total: 8 roles in one quarter and none in the next two argues against a full-time hire.
Can you use recruiting software and an agency at the same time?
Yes, and it is the most common sensible setup. Run every candidate through your own system regardless of who sourced them, so agency-sourced people, referrals and inbound applicants all end up in one pipeline with one evaluation standard and one history that stays with you when the contract ends.
Free recruiting audit · 2 minutes
Find out what your hiring process is actually costing you.
Answer eight questions, or connect your current system read-only, and get a report on where your funnel loses candidates and which changes are worth making. No signup, no API key stored, data stays in the EU.
Written by Andreas Amann
Founder of Pickr. Former operator at startups in Berlin and Silicon Valley, where he helped scale companies from 40 to 200+ people. Built Pickr after years of using every major ATS as a recruitment agency owner at ScalingPPL.