The fee is a percentage of first-year salary, payable when the person starts, with a replacement guarantee if it falls apart inside three months. Your role has been open seven weeks. You have read ninety applications and phone-screened five people, two of whom did not show up. Now you have to decide whether paying an agency is a sensible expense or an admission that you cannot hire.
Neither. It is a pricing question with an answer you can calculate.
A staffing agency fee is worth paying when the cost of your own hours plus the cost of the role sitting empty exceeds the fee, and when the role is genuinely hard to source. Two tests, and you need both. Fail the second and you are paying a premium for candidates who would have answered your own advert. Fail the first and you are paying to save time that was not costing you anything.
Most owners asking is a staffing agency worth it compare the fee against zero, because a self-run search feels free. It is not. It is unbilled.
What the fee actually buys, which is not candidates
You think you are buying a person. Every candidate an agency sends exists whether or not you pay, and some would have found your posting anyway. You are buying three other things.
Search time that is not yours. Someone else writes the outreach, chases non-responders, handles the candidate who goes quiet for four days, runs the first conversation that eliminates most of the field. That work happens either way. The only question is whose calendar it happens on.
A shortlist filtered by someone who has seen the market. A recruiter who has placed eleven dental office managers in your metro this year knows what those people earn, who is unhappy where and which of your requirements is killing your applicant flow. You know what your last office manager was like. Sample size of one.
The option to say no cheaply. If the agency’s four candidates are all wrong, you have spent two hours. If your own search produces four wrong candidates, you have spent six weeks producing them and you restart from a cold advert. On contingency you pay only on a hire, so the failed search is their loss.
The break-even, done properly
Substitute your own numbers. Mine are illustrative except where flagged.
Take a bookkeeper. The Bureau of Labor Statistics puts median pay for bookkeeping, accounting and auditing clerks at $50,670 as of May 2025, so call the job $55,000. That is not what the role costs you: BLS Employer Costs for Employee Compensation for June 2026 puts benefits at 30.0% of total compensation for private industry workers, so loaded cost is nearer $78,600.
Side one: your search hours, for one search that works first time. Writing and placing the posting, 3. Reading 120 applications at two minutes each, 4. Twelve phone screens at 25 minutes including the three who do not answer, 5. Scheduling and chasing, 3. Six first interviews plus notes, 5. Three second interviews, references, offer call and a negotiation, 6. Twenty-six hours.
Now price your hour. Not salary divided by 2,080. Your hour is worth whatever the best alternative use of it produces: gross profit per selling hour if you are the one who closes, your billable rate if you are the delivery bottleneck. Say $150, so 26 hours is $3,900.
Side two: the empty seat, per week. Work not getting done, plus work landing on someone who should be doing something else. If the vacancy means you personally spend eight hours a week on reconciliations at that $150, that is $1,200 a week, before late invoicing, cover overtime and the customer who waited. For a revenue-touching role this dwarfs everything else here. For a back-office role it may be small, and you should say so rather than inflate it.
So the comparison is not “fee versus my time.” It is:
Fee versus (your hourly value × hours the agency actually saves) + (weekly cost of vacancy × weeks the agency actually saves).
Both “actually” clauses do work. An agency does not save all 26 hours; you still interview, decide and negotiate. It takes the sourcing, screening and scheduling load, the first four lines above, around 15 hours. So $2,250, not $3,900.
Weeks saved is the bigger lever and the harder estimate. No trustworthy published time-to-hire average applies to your role in your town, and anyone quoting one has borrowed it from a vendor survey. Bound it instead: ask how long their last three placements in this role family took from brief to start date, role by role rather than averaged, then assume you are seeing their best work.
Put three weeks on it at $1,200: $3,600. Add the $2,250 and your break-even fee is around $5,850 for this hire, before any quality benefit. Under that, defensible. Well over, and you are paying for something else, which you should be able to name.
Ask for the fee as a dollar figure on this salary, in writing, and ask what it is calculated on: base only, or base plus commission, bonus and sign-on. Percentage framing hides the size of the cheque, and “first-year cash compensation” is a bigger base than “salary” wherever there is variable pay.
A bad hire and a slow hire cost different things
You will see a claim that a bad hire costs 30% of first-year earnings, usually credited to the US Department of Labor. Treat it as folklore; the attribution travels without a citation, and one percentage cannot describe both a mis-hired warehouse picker and a mis-hired controller. Build your own from five parts: wages and benefits paid while you worked out it was not working, your supervision hours at your hourly value, rework and client damage repaired, the second search at whatever the first cost, and the load on people who covered twice.
An agency reliably reduces the cost of a slow hire. It reduces the cost of a bad hire only partly, and only through the mechanism written into your contract. A recruiter cannot know whether someone will thrive in your business; they have never watched your supervisor manage anyone. They can eliminate candidates who fail on verifiable facts and keep you from interviewing six people who were never going to work, which raises the quality of the field you choose from. Real, but partial.
The replacement guarantee is the actual protection, and three things decide its worth. Replacement search or cash refund, and if replacement, on what timeline? Does it cover a redundancy, or only resignation and dismissal for cause? Does it lapse if you pay the invoice late, which is standard and easily missed? A replacement-only guarantee with late-payment forfeiture is much weaker than the same period with a pro-rata refund.
When a staffing agency is worth the fee
Hard-to-source skills. The people who can do the job are employed, not looking, and will never see your advert. Skilled trades, licensed clinical roles, niche engineering, any discipline where good practitioners move through conversations rather than applications. Adverts do not reach passive candidates. Outreach does, and it is labour you will not do well between running a business.
A confidential replacement. The incumbent still works for you, or the role’s existence signals something to staff or competitors. You cannot run that search yourself without it leaking. An agency approaches the market without your name attached until you choose to attach it. No self-service substitute exists here.
A role you have already failed to fill once. The strongest signal here, and the most ignored. A failed search told you something: salary off market, requirements unrealistic, job title not what candidates search for, or supply genuinely absent. A recruiter in that role family can usually say which, often on the first call, and the diagnosis is worth having even if you then run the search yourself.
Temporary and seasonal cover. A stronger proposition, because you are buying an employer of record rather than search efficiency: the agency payrolls the worker, carries employment tax and workers’ compensation, and absorbs the end of the assignment. American Staffing Association member-reported data puts nearly 2.2 million temporary and contract employees working for US staffing firms in an average week in 2024, so this is routine. One warning: the payroll moves, your safety duty does not. OSHA’s position is that the agency and the host employer “are jointly responsible for maintaining a safe work environment for temporary workers,” and site-specific hazard training is normally the host’s job. That is you.
Any hire where your time is the binding constraint. If you are the only person who can sell, close, treat patients or ship the work, and searching competes directly with that, the arithmetic is not close.
When it clearly is not
Abundant local supply, proven by your own funnel. An empirical test. Post the role for ten days and count qualified applicants, not applicants. Eight people you would genuinely interview means you have no sourcing problem and the agency is selling you something you already have. Checking beats assuming: BLS JOLTS reported 7.3 million job openings against 5.1 million hires in July 2026, a market where plenty of roles fill and plenty stay open, which tells you nothing about yours. Your applicant count does.
Junior roles inside your own network. Where the qualification is attitude plus trainability, your staff already know people. A referral bonus of a few hundred dollars paid at 90 days costs a fraction of a placement fee and comes with a built-in reference.
Anything where you cannot describe what good looks like. The disqualifier, and no fee fixes it. If you cannot write down what this person will have accomplished in six months, what they must do unsupervised on day one, and which requirement you would waive for the right person, you will reject every shortlist for reasons you discover only on seeing the CV. You will burn the agency’s goodwill, get the B-team recruiter next time, and conclude agencies do not work. The brief is the product. Fix it first. Same goes if you are unsure the role should exist or the budget is not signed off.
Fee structures and the four things worth negotiating
General information, not legal advice. Recruitment fee agreements are commercial contracts governed by state law, and terms like exclusivity, rebate periods, indemnities and non-solicitation clauses carry consequences specific to your jurisdiction. Have anything you are about to sign read by someone qualified.
No percentages here. Fees move with role, seniority, geography, exclusivity and volume, nobody publishes them, and a number in an article becomes an anchor that has nothing to do with your market. Learn the mechanisms.
Contingency versus retained. Contingency pays only on a hire: low risk for you, and a weak claim on the recruiter’s calendar, since they run several searches and prioritise the ones likeliest to close. Retained pays in instalments regardless of outcome, usually engagement, shortlist and placement, buying committed effort plus exclusivity. For most small business hires, contingency. Retained earns its place on a scarce senior role you need searched properly rather than skimmed.
Contract and temp markups. Not a fee, a multiplier: bill rate equals pay rate plus a markup covering employer tax, workers’ compensation, insurance, overhead and margin. Ask what the worker actually receives, what the markup includes, and what happens on overtime, holidays and statutory wage increases. That last mechanic is where quoted rates and invoiced rates part company.
Temp-to-perm conversion. Negotiate before the assignment starts, never during. The fee should decline with hours worked to a stated zero point, and “conversion” should be defined to include hiring the person through another agency or an affiliate. This route costs more in cash than a straight placement over a few months, and buys you something else: you watch the person do the real job before taking on the employment risk.
Guarantees, rebate periods and exclusivity. Guarantees, above. Agencies discount for exclusivity and it can be a fair trade, but only with a time limit and a performance condition: exclusive for 21 days, first shortlist inside seven, reverting automatically. Never grant open-ended exclusivity on a role you urgently need filled. Agree a source-priority rule in writing too, covering who owns a candidate already in your records or applying directly. That clause prevents the most common fee dispute in small business recruiting.
How to brief an agency so it works
Most small business engagements fail here, not at the fee. The recruiter takes a twenty-minute phone brief, sends four CVs, hears nothing for nine days, and places their best candidate elsewhere. You concluded the agency was useless. They concluded you were not a real client. Both were right. Four things fix it, and they take an hour.
Write a one-page scorecard. Not a job description. Four or five outcomes with numbers and dates: close the month within five working days by month two, clear the 60-day-plus receivables backlog by quarter end. Outcomes give a recruiter something real to screen against, and let you compare candidates on the same axis instead of on likeability.
Separate must-haves from nice-to-haves, and keep must-haves to three. Nine requirements tell a recruiter nothing about which ones you mean, so they screen conservatively and you get over-qualified people who want more than you have. Say which requirement you would waive for the right person, and which you never would.
Be honest about salary, ceiling included. A recruiter paid on a percentage of salary has an obvious incentive to push your number up, which is exactly why you give a real range and a real maximum rather than a fake low one you will abandon. Understating it produces candidates who withdraw at offer stage, costing you weeks and them the fee.
Commit to a turnaround out loud. Feedback on every submitted CV within 48 hours, even if it is one line. Interview slots held in advance rather than found after the shortlist lands. A decision within a stated number of days of the final interview. Where good candidates sit in several processes, the slowest employer loses, and your one advantage over a larger competitor is that you can decide on Thursday instead of convening a panel in three weeks.
Good recruiter or CV forwarder
A CV forwarder runs keyword searches on a database and sends volume, hoping something sticks. A recruiter works a labour pool and knows who is in it. The difference shows in fifteen minutes.
- They push back on your brief, telling you the salary is under market or that your combination of skills barely exists locally. Agreement with everything you say is a sales posture.
- They ask about the manager, the team and why the last person left, because they must sell your job to someone who has options, and a job description will not do that.
- They talk about people, not process: “maybe forty people do this in this metro, I have placed six, two I would call today.”
- They describe the market unprompted: what these roles pay now, what shifted this year, which local employer just froze hiring and released good people.
Ask, more or less verbatim:
- “How many roles like this have you filled in the last twelve months, and where?”
- “Who actually works my search, you or someone else? Can I meet them?”
- “How will you source this: database, job board or direct approach, in what proportion?”
- “Talk me through a search you failed on and why.” An honest answer is the best signal on the call.
- “Before you send anyone, what would you change about my brief?”
- “How many candidates will I see?” Four well-argued profiles beat fifteen; volume hands the screening back to you while charging as though they had done it.
- “If four weeks in you decide this role is unfillable at my budget, what happens?” You want to hear that they would tell you and stop. Contingency economics reward staying quiet and sending more CVs.
Questions people ask next
Should I give the role to two or three agencies at once? Two, sometimes. Three is self-defeating: each recruiter sees a one-in-three chance at best, deprioritises you, and sends whatever is nearest to hand. With two, tell both, agree in writing who owns a duplicate candidate, and give the same brief and the same deadline. On fees, negotiate by trading rather than asking: time-limited exclusivity, a second hire, faster payment terms. A discount extracted for nothing gets recovered in the attention your search receives.
Are agencies interested in lower-paid roles? Many are. ASA’s member-reported occupational mix puts roughly 36% of staffing employees in industrial work and 24% in office-clerical and administrative, against 21% professional-managerial. An agency that says your role is too small is describing its own model, not your options.
AB7 Solutions does exactly the work this article describes. If your time is the binding constraint, we take the search: recruitment and RPO run against a scorecard rather than a keyword sweep, plus contract staffing, C2C and remote professionals where the honest answer is flexible capacity instead of a permanent head, with payroll and employer administration on our side. We will also tell you when you do not need us. Send us the role, the salary and what your own advert has produced so far. If your funnel is healthy and what you need is a sharper brief and two afternoons of interviews, that is what you will hear; if it is not, we will name which of the usual causes is breaking your search before you spend anything. Call +1 321 341 7733, email ab@ab7solutions.com or director@ab7solutions.com, or start at www.ab7solutions.com.
Sources: U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, Bookkeeping, Accounting, and Auditing Clerks (median pay, May 2025), Employer Costs for Employee Compensation, June 2026, and Job Openings and Labor Turnover, July 2026; American Staffing Association, Staffing Industry Statistics (member-reported survey data); OSHA, Protecting Temporary Workers. Fees, markups, hourly values and search hours in the worked example are illustrative, not published figures.