The September jobs report looked calm, and that is exactly the problem for anyone carrying an open technical requisition. The Bureau of Labor Statistics reported on October 2 that nonfarm payrolls rose by just 29,000 in September, the unemployment rate was little changed at 4.2 percent, and July and August were revised down by a combined 60,000 jobs. A week earlier, the BLS JOLTS release showed 7.1 million job openings in August, the lowest level since March, against 5.2 million hires, 3.1 million quits and 1.6 million layoffs and discharges. Employers are not firing. They are also not hiring. That combination turns every unfilled role into a slow, compounding cost.
What did this week's labor data actually say?
Three numbers matter for CIOs and talent leaders. First, hiring momentum is thin: September's 29,000 payroll gain sits well below the 45,000 average monthly gain of the prior twelve months, and the summer was weaker than first reported once revisions landed. Second, the openings-to-hires gap is still wide. With 7.1 million openings and 5.2 million hires in August, roughly 1.9 million more roles were posted than were filled that month, so approved requisitions are competing for a limited flow of completed hires. Third, the growth that did happen is relevant to technical teams: manufacturing added 9,000 jobs in September and 72,000 since its December 2025 low, and construction added 11,000. The roles staying open longest in operating companies are the unglamorous, business-critical ones: firmware and embedded engineers, test and validation specialists, network and infrastructure staff, cybersecurity analysts, SRE and NOC operators, field-service technicians and ERP support. Low layoffs mean those people are employed. Low quits, at 3.1 million, mean they are not moving. Your vacancy is not waiting for a wave of applicants. It has to go and find them.
Why does a slow market make each vacancy more expensive, not less?
It is tempting to treat an unfilled role as money saved. For a payroll line, it is. For an operating team, the cost just changes shape. The project the role was hired for still has a date. The systems the role was hired to run still break, patch and audit. The work lands on the people you already have, in overtime, in context-switching and in quiet deferral of everything that is not on fire this week. BLS put average hourly earnings for private nonfarm employees at 37.81 dollars in September, with an average workweek of 34.4 hours. Use that only as an anchor for your own arithmetic: at roughly 260 dollars of wages per working day for an average employee, a single role left open for 60 working days represents on the order of 15,000 to 16,000 dollars of missing paid capacity, before overtime premiums, contractor cover, delayed revenue or a slipped launch enter the calculation. For a senior firmware engineer, a security analyst or a field-service lead, your real number will be higher, and you should calculate it with your own salary band rather than a national average. The point is not the exact figure. The point is that the figure grows every week, whether anyone writes it down or not.
How do you calculate your own cost of vacancy?
Price the role weekly, in four lines. Line one is missing capacity: the role's loaded daily cost multiplied by working days open. Line two is absorption cost: overtime paid, contractor or vendor cover bought, and the measurable time senior staff spend doing work below their level. Line three is delay cost: the launch, migration, audit or customer commitment that moves because this seat is empty, stated as a date and, where you can, a revenue or penalty figure. Line four is restart cost, the one most teams forget: every week a search sits idle, the candidate pool you already found goes stale, referrals cool and the search effectively starts again. Add the four lines and put the total in the same review where the requisition itself is discussed. A vacancy with a weekly price behaves differently in a meeting from a vacancy that is just an open headcount. It gets staffed, re-scoped or honestly killed, and all three are better outcomes than drifting. The tax also compounds across a cluster, not just a single seat. A test engineer, a field-service lead and an ERP support specialist open at the same site for the same quarter do not create three separate inconveniences; they create one operating problem, where senior staff triage tickets instead of shipping, installations wait on the same missing hands, and every delay is used to justify the next one. Price the cluster together and the conversation changes from "recruiting is slow" to a specific weekly figure an operations leader can act on.
Where do CIOs and talent leaders cut the cost first?
Not by lowering the bar, and not by posting the same job in more places. Cut time-to-evidence. Write the requisition around proof you can score: the systems, certifications, environments and shipped work this person must actually have, for the roles above, in your industry, in the United States or Canada. Then widen the search beyond the inbound funnel, because in a market with 3.1 million quits and falling openings, the right candidate is employed and not applying anywhere. That is the work UPPER is built for: sourcing channels worked systematically, every candidate scored against the same evidence, and a ranked shortlist your own recruiters operate and your hiring manager can judge quickly. The economics are simple. If a role costs your team real money every week it stays open, then a search that reaches a judgeable shortlist in days rather than months is not a recruiting upgrade. It is cost control, in the same category as renegotiating a vendor contract, and it should be measured the same way.
What should you measure on Monday morning?
Start with five numbers per open technical role: days open, the four-line weekly vacancy cost, how many candidates have been scored against written evidence, how many the hiring manager has actually judged, and the date of the next decision. If any role cannot produce those five numbers, the search is not being managed, it is being waited on. September's data says waiting is a strategy the market will punish quietly: few layoffs to free up talent, few quits to circulate it, and 7.1 million openings competing for 5.2 million hires. The teams that come out of a low-hire market ahead are not the ones that hired the most. They are the ones that knew what each empty seat cost, and refused to pay the vacancy tax one week longer than they had to.