On September 3, ADP released its National Employment Report for August. Private payrolls: 54,000. Weak. A second consecutive miss well below the approximately 150,000 monthly additions the U.S. economy needs to keep unemployment from drifting up.
Twenty-four hours later, the Bureau of Labor Statistics released the August Employment Situation. Total nonfarm payrolls: 162,000. A three-times beat against the roughly 53,000 consensus. The strongest print in six months.
Same month. Same economy. A gap of 108,000 jobs depending on which report you read.
If you are a recruiter going into Q4 conversations and a hiring manager asks you whether the labor market is "strong" or "soft," you need to be able to answer that question without using either number as a crutch. Because both are correct, and the difference between them tells you more than either headline does on its own.
Why These Two Reports Measure Different Things
ADP processes payrolls for a large share of U.S. private businesses. Its National Employment Report draws on actual payroll records (paychecks cut during the reference period) and tracks month-over-month changes across that client base. It covers private employment only. Government workers do not appear in ADP's count.
The BLS Employment Situation is a survey. Each month, the Current Employment Statistics program surveys approximately 130,000 business establishments and government agencies, representing roughly 670,000 individual worksites. That sample is then weighted to produce a national estimate of total nonfarm payroll employment, private and public combined. The BLS number is the one the Federal Reserve, the White House, and financial media treat as the official read.
Both methodologies have known strengths. ADP is cash-register data: when a payroll gets processed, the job is real. BLS has broader coverage including government hiring that ADP skips. Neither is wrong. But they answer slightly different questions, and in August 2026, the difference between those questions was 108,000 jobs.
What Explains the Gap This Month
The August BLS number included two sector gains that ADP would not capture in the same way.
Government education added 42,000 jobs in August, according to BLS. This is largely a seasonal reversal: August schools came back online after a July that saw a 50,000-job loss that was itself a calendar artifact. The BLS seasonal adjustment model for education has struggled with varying school-calendar schedules across districts. The August education gain is real in the sense that school employees returned to work. It is not new demand. And it does not appear in ADP's private payroll count.
Food services and drinking places added 59,000 jobs, according to BLS, nearly five times the sector's 12-month average of 12,000. This spike is concentrated in private employment, so ADP does pick it up. But ADP's own methodology tracks a narrower subset of employers, and timing differences in payroll processing cycles can create a one-month lag in how rapidly-growing seasonal sectors show up in ADP versus BLS.
Strip out the 42,000 government education jobs and the 59,000 food service spike, and the rest of the August BLS data comes in at approximately 61,000 jobs, still stronger than ADP's 54,000 but not dramatically so. The two datasets, when you read them at the sector level rather than the headline, are describing a similar underlying economy: modest private sector growth, concentrated in specific pockets, not broad-based.
That alignment is the signal. The divergence in the top-line numbers is almost entirely methodological, not economic.
Which Number Should Drive Your Q4 Conversations
For recruiters, the honest answer is: use both, for different purposes.
The BLS 162,000 headline is the right number for hiring manager conversations about whether the market is "improving." It captures everything (government, education, seasonal hospitality) and it represents the broadest authoritative read on U.S. employment conditions. If a client has been sitting on headcount approvals because of July's -23,000 print (since revised to +21,000) and needs a reason to reopen requisitions, the BLS number gives you that reason. The labor market did not contract this summer. It stalled, then recovered. That is the correct framing.
The ADP 54,000 private payroll figure is the right number for setting expectations about sourcing speed and candidate availability in the private sector. If your clients are in professional services, finance, technology, or manufacturing, they are competing for candidates in a private-sector hiring environment that has run below sustainable levels for two consecutive months. Passive candidates are not moving (the JOLTS July quit rate of 1.9% is near a multi-year low, per BLS). The private-sector pool of actively-searching candidates is not growing quickly.
Using BLS 162,000 to tell a tech or finance hiring manager "the market is hot, act now" is not wrong in spirit but it is misleading in execution. The heat in August was concentrated in food service and school calendars. Their sector is competing in the ADP economy, not the BLS economy.
The Revision Problem, Applied Forward
There is one more layer to this. BLS monthly estimates are published quickly and revised repeatedly. July's initial -23,000 print was revised to +21,000 when August's data dropped, per BLS. June went from +20,000 to +31,000.
ADP numbers are not revised in the same way. The company re-benchmarks its methodology periodically, but monthly figures do not get retroactively updated the way BLS establishment survey data does. This means ADP gives you a more stable read on a narrower universe, while BLS gives you a broader read on a more volatile one.
For Q4 planning conversations, this matters. The narrative around the labor market has shifted twice in six weeks: from "July went negative" (original BLS) to "August smashed expectations" (revised July + strong August). Hiring managers who made decisions based on the July negative print made them on data that was later corrected by 44,000 jobs.
If you are advising a client who paused hiring in late August on the strength of the July -23,000 headline, walk them through the revision. The contraction scenario that justified the pause no longer exists in the data.
How to Run the Q4 Conversation Right Now
The clearest framing for a hiring manager in September 2026:
The labor market is not contracting. August's 162,000 BLS total, combined with the revision of July's -23,000 to +21,000, makes that clear. Confidence is warranted for Q4 headcount.
The labor market is also not overheating in a way that should change what you offer candidates. Average hourly earnings rose 0.3% in August, or 3.1% year-over-year, per BLS, a stable pace that tracks inflation rather than one signaling wage escalation. Candidates in the private sector are not fielding five competing offers. The ADP data confirms this.
ADP and BLS together tell you: the window is open, and the competition is manageable. That combination does not last. Use it.
If you want a single view of real-time demand signals and candidate availability by sector, BlueLine is free to try at /register.