Sizing a staffing niche with labor statistics, posting volume, fee norms and competitor density, plus the signals that should send you somewhere else entirely.

Part One: Why Your Gut Makes a Lousy Market Analyst

1. Why does almost every new staffing firm pick its niche the same way?

Because the hunch is free, fast and flattering. Your old industry hands you a vocabulary, a contact list, and the warm feeling that you already know something. That is real. The problem is that your resume answers exactly one question, whether you can hold a conversation with a hiring manager in that space, and leaves six others untouched. Jason Cohen, who bootstrapped Smart Bear and then built WP Engine, described the death spiral on A Smart Bear: a founder gets a flash of insight, talks to three people who confirm the problem is real, builds a solution that works, and shuts the doors in 6 to 24 months wondering what happened. The req is real. The hiring manager nods. The desk still starves.

2. How often does the hunch approach actually blow up?

Often enough that it shows up plainly in the data. StaffingHub’s 2026 State of Staffing Benchmarking Report, built on 231 responses from executive, VP, director and manager level operators, found 42% of agencies lost revenue in 2025, one in five of those by more than 30%. Healthcare staffing took the worst beating, contracting 56% against 39% across other verticals. Pair that with Failory, which has spent years interviewing founders about why their companies died: across 80-plus failed startups, marketing problems led at 56%, team at 18% and finance at 16%. That is a demand-side diagnosis. These people did not fail at building. They failed at picking the room.

3. What does the 2026 market look like for somebody about to plant a flag?

Flat on the surface, wildly uneven underneath. Indeed Hiring Lab’s Job Postings Index sat at 101.8 on August 14, 2026, just 1.8% above the February 2020 baseline after peaking near 160 in early 2022. About half of all tracked sectors now sit at or below pre-pandemic levels. The spread is the story: many engineering and healthcare fields run roughly 30% above baseline while marketing, data analysis and software development have fallen about 30% or more. Software development registered 74.4 in August 2026, up from a floor of 61.1 in May 2025. Posted wage growth has slowed to 2.4% against CPI inflation of 3.5%, so clients are stingier with markups. The headline number tells you almost nothing. The dispersion tells you everything.

Part Two: Size the Pool Before You Name the Company

4. What is the actual arithmetic for sizing a vertical?

Steal Cohen’s Fermi rubric and point it at staffing. He argues you need roughly 10 million consumers or 100,000 businesses with the problem, because a strong ad campaign converts about 1% of impressions to visitors and a good site converts about 1% of those to buyers. That is 10,000 impressions per customer. His scorecard multiplies seven factors: Plausible, Self-Aware, Lucrative, Liquid, Eager by identity, Eager by differentiation, and Enduring. Multiply, divide by 625,000. A score of 1 clears the bar for a small independent firm. A 2 or better can scale.

5. How does that translate to a staffing desk?

Swap “customers” for “hiring managers with recurring reqs,” then run the vertical through the same gates:

•  Plausible: how many employers in your geography hire this title at all? Powers of ten only. 1,000 is a hobby, 10,000 is a desk.

•  Self-Aware: do they already use agencies, or would you be inventing the category? Score 0.01 if few care, 0.5 if it is standard practice.

•  Lucrative: annual allocated external recruiting budget per account, rounded to a power of ten.

•  Liquid: 0.01 if they hire this role every few years, 1.0 if they are always in the market. This line kills more staffing verticals than any other.

•  Enduring: 0.5 for recurring perm hiring, 1.0 if you become the system of record for a workforce that redeploys.

6. Where do I get labor numbers that are better than a Google guess?

Revelio Labs publishes Public Labor Statistics from over 100 million US profiles covering roughly two-thirds of employed people, against an estimated 27% for the BLS establishment survey and 0.03% for the household survey. In August 2026 it showed 36.5 thousand jobs added, active postings down 3.0% for the month, and the hiring rate slipping from 20.1% to 19.7%. That last number is a gift, because the hiring rate is your addressable churn. A vertical at 25% has four times the natural order flow of one at 6%, before you make a single call.

Part Three: Posting Volume Is a Pulse, Not a Verdict

7. Is a mountain of job postings good news or a warning?

Depends entirely on what is in the pile. Indeed Hiring Lab found senior-level postings surged almost 15% between May 2025 and May 2026 while entry-level postings fell 6.2% from their January 2025 level. The Burning Glass Institute documented entry-level postings dropping roughly 37% between 2022 and 2024, and Revelio Labs pegs the slide at about 35% since early 2023. Two organizations, two methodologies, same verdict. If your thesis rests on placing juniors at volume, you are betting against a trend two separate datasets agree on.

8. What can the language of the postings themselves tell me?

More than most founders think. Datapeople has analyzed over 100 million job listings alongside millions of hiring outcomes and found posts padded with “preferred” qualifications draw 40% smaller qualified candidate pools and 40% smaller female applicant pools than posts listing only must-haves. They also found 86% of tech applicants arrive through organic sources. Read that as a market signal. If postings in your vertical are bloated, vague and pulling in no one usable, the client is already in pain and does not know why. That is a live account. Also check the clock you inherit: Bullhorn GRID 2026 data shows 56% of top performers fill in under 10 days, and an elite 22% fill in under 3.

Part Four: Do the Fee Math Out Loud, On Paper

9. What margin should I plan for?

Less than the deck in your head says. StaffingHub found median gross margin between 20% and 29% across contracting, flat and growth firms alike, so margin did not separate winners from losers. Net profit commonly lands at 3% to 6% once overhead, compliance and benefits burden take their cut. Segment margins vary: roughly 26% in IT, 38% in finance and accounting, 21% in healthcare. Learn the denominator trick before a client uses it on you. A $30 bill rate against a $22 pay rate is an $8 spread, which is 27% as margin and about 36% as markup. Same eight dollars, and your salespeople and your accountant will never agree unless you name which one you mean.

10. Can you show the fee math on a real vertical?

Take travel nursing, where BluePipes has published the mechanics for years. A standard contract runs three 12-hour shifts a week for 13 weeks, or 468 hours. At a $90 bill rate it bills $42,120. Apply a 21% healthcare gross margin and you clear roughly $8,845 per placement. Now the sobering part. Industry aggregates put top-quartile recruiter gross profit near $380,000 a year, so reaching it takes about 43 contracts per recruiter. Write that number down before you register the LLC.

11. How many placements does a normal human actually make?

This is where the fantasy dies quietly. Recruiterflow’s benchmark work shows the top 25% of recruiters place 5.21 candidates a year while everyone else places 1.38. The gap opens widest at one stage: top-quartile firms convert 50.1% of screens into submissions against 36.1% for the rest. And 71% of placements come from candidates already sitting in the CRM. So favor markets where the same candidates recycle. A niche where every placement demands a brand new search is one where you never build an asset.

Part Five: Competitor Density Is Not a Headcount

12. How do I measure how crowded a vertical really is?

Counting logos on LinkedIn tells you almost nothing. Ask who owns the distribution. StaffingHub found growth agencies twice as likely as flat or contracting ones to run an all-owned top-three sourcing mix, 50% versus 27%, while job boards made the top three for only 44% of growth agencies against 73% of the rest. The density that matters is rented-channel density. If every firm in your vertical bids for the same Indeed traffic, you are entering an auction.

13. What separates the firms that grew from the ones that shrank?

Three things, and pricing was not one of them. StaffingHub scored operational maturity on a seven-point scale (weekly KPI review, weekly pipeline meeting, a named scorecard owner, documented SOPs, conversion tracking). Growth agencies averaged 4.56, contracting agencies 3.56. Heavy AI adopters using five or more processes grew 39% in 2025 while agencies using no AI grew 17%, a 2.3x difference. And 62% of fast-growth agencies planned new software purchases against 37% of slower firms. Depth beat presence. One chatbot does not count.

Part Six: The Disqualifying Signals

14. What should make me walk away from a vertical entirely? These do not get fixed by working harder.

•  Liquid score of 0.01. The client makes this hire every three years. Episodic demand means episodic revenue.

•  No allocated budget. Cohen watched 17 years of Capital Factory startups try to sell to college students. Real problems, zero winners, because the buyer had no money. The staffing version is a vertical where every employer insists on hiring direct.

•  A DSO that outruns your payroll. Healthcare staffing commonly runs 45 to 60-plus days because hospital revenue cycle management is slow, while a diversified public firm implies roughly 49. If DSO exceeds your payroll cycle by more than 30 days, you need factoring before you take the first order.

•  Entry-level heavy demand, with postings in that band down roughly 35% to 37% depending on the dataset.

•  An internal team already tasked with the problem. You want buyers who already outsource, because internal teams fight vendors that make their jobs look optional.

Part Seven: Three Examples From the Future

15. Show me the scorecard doing real work.

Priya, January 2027, industrial maintenance electricians in the Ohio Valley. Production and Manufacturing postings ran about 8% above baseline through 2026. Plausible: 10,000 employers. Self-Aware: 0.5, contract trades labor is standard practice. Lucrative: $10,000 a year. Liquid: 1.0, they are always short. Eager identity 0.5, Eager comparative 0.5, Enduring 1.0 because contractors redeploy. Multiply and you get 125,000,000, which divided by 625,000 is a score of 200. Not a typo. Recurring contract labor in a tight trade produces absurd scores because Liquid and Enduring both land at 1.0.

16. Now show me one that fails.

Marcus, March 2028, builds an AI/ML search firm because the headlines told him to. Plausible: 10,000 companies hiring at that level. Self-Aware: 1.0, they absolutely know. Lucrative: $100,000 in search budget. Liquid: 0.1. Eager identity 0.1, since every retained firm in the country arrived first. Eager comparative 0.1. Enduring 0.1. The product is 1,000,000, which divided by 625,000 gives 1.6. Viable on paper, brutal in practice, and all the weakness sits in the differentiation and durability lines. The market is enormous and he is invisible in it.

17. And the one that no sane person would pick on a hunch?

Dana, September 2029, allied health in secondary metros: respiratory therapists and surgical techs in cities under 500,000. Healthcare is about 11% of US employment but delivered nearly three quarters of all net job growth in 2025, and employers keep saying they cannot find qualified workers as the population ages. Dana scores Plausible at just 1,000 employers, which looks fatal, but Lucrative jumps to $100,000 per account and Eager identity climbs to 1.0 because she is the only specialist who returns calls in Youngstown. Small pool, enormous per-account value. That is the whole point of niching: the customer count drops and four other numbers go up.

Part Eight: The Twenty-Minute Version

18. Compress it for me.

•  Pull the sector posting index and the sector hiring rate. Below baseline and falling means you need a reason that is not nostalgia. The hiring rate is your natural order flow; 19.7% was the US average in August 2026.

•  Score the seven Fermi lines in powers of ten, multiply, divide by 625,000. Under 1 means find another vertical.

•  Compute gross profit per placement at a realistic segment margin, 21% to 38%. Divide $380,000 by it. That is your annual quota.

•  Check the DSO norm against your payroll cycle; a gap over 30 days is a financing problem. Then check whether incumbents rent their channel or own it. Rented means auction, owned means moat.

19. Last thing. What is the single most expensive mistake here?

Treating a vertical choice as an identity instead of a hypothesis. You are not swearing an oath to industrial electricians. You are making a falsifiable bet with an entry price, a break-even placement count, and conditions that tell you to quit. Failory’s founder interviews keep landing on the same finding: the failure is demand-side, and it was visible before the money got spent. Two weeks of arithmetic beats two years of conviction.

Sources Cited

•  Indeed Hiring Lab, US Labor Market Snapshot, August 2026  https://www.hiringlab.org/

•  Indeed Hiring Lab, Labor Market Tilting Toward Seniority  https://hiringlab.indeed.com/

•  StaffingHub, 2026 State of Staffing Report  https://staffinghub.com/state-of-staffing/

•  Revelio Labs, Public Labor Statistics  https://www.reveliolabs.com/blog

•  Burning Glass Institute, research library  https://www.burningglassinstitute.org/research

•  Datapeople, Best-in-Class Qualifications  https://datapeople.io/blog/

•  Recruiterflow, 25 Recruitment Metrics for 2026  https://recruiterflow.com/blog/recruiting-metrics/

•  Loxo, Most Important Recruiting Metrics  https://www.loxo.co/blog/

•  BluePipes, How Agencies Calculate Pay Rates  https://blog.bluepipes.com/

•  A Smart Bear, Excuse me, is there a problem?  https://longform.asmartbear.com/problem/

•  Failory, Startup Failure Rate  https://www.failory.com/blog/startup-failure-rate

•  Aptitude Research, Hiring Automation 2026  https://www.aptituderesearch.com/