Kelly Finally Grew — Just Not Where It Promised

Seven quarters, one number

Kelly Services spent most of 2026 asking investors to look through a revenue line that kept shrinking and grade it on the "underlying" number — the business stripped of the federal-government and big-client step-downs that management had flagged as discrete. In Q3 that request finally paid off. Revenue of $927 million fell 0.8% year over year, but the number the company wanted graded, underlying revenue, “grew 2.3%, ahead of our guidance of 1% to 2%” — Troy Anderson, Chief Financial Officer · 2026-11-05 — a 290 basis point sequential improvement that, by Troy Anderson's math, supplied more than half of the total year-over-year improvement in the quarter.

The engine was Enterprise Talent Management, the segment that has been the company's problem child since it was cobbled together. Underlying ETM revenue rose 7.7%, with Talent Solutions up 11.6% on double-digit MSP and RPO growth and underlying staffing up roughly 7%. Chief executive Chris Layden did not bury the milestone: it was “the first time that we've seen year-over-year growth since the segment was established, that's 7 quarters out” — Chris Layden, Chief Executive Officer · 2026-11-05. SET also improved sequentially for a second straight quarter, with higher-margin solutions now 40% of that segment. Adjusted EBITDA of $21.3 million rose 29% on lower revenue and margin expanded 50 basis points to 2.3% — the first year-over-year margin expansion in at least four quarters.

The catch is the base. Total revenue in the last filed quarter was $1.0 billion, 23% below where it sat eleven years ago. A decade of mix shift toward solutions bought three points of gross margin — gross margin now sits at 20.4% — while operating margin stayed stuck near 1.6%. Kelly has become a better-margin business without becoming a more profitable one.

The Education promise, deferred again

Education is where the quarter broke. Revenue fell 6.4%, and the CFO was blunt: “Education revenue declined 6.4%, which was below our expectations” — Troy Anderson, Chief Financial Officer · 2026-11-05. The cause is a real-world variable Kelly does not control — enrollment pressure in K-12 districts, which Layden described as “near-term enrollment pressure in some K-12 districts, which became more visible as we began a new school year in the quarter” — Chris Layden, Chief Executive Officer · 2026-11-05. Fewer students means fewer classrooms, which means less demand for substitute teachers — even after what management called a strong selling cycle.

What makes this a story rather than a data point is the trail behind it. In May, Layden told investors the Florida enrollment hit was behind the company and “that demand has now stabilized” — Chris Layden, Chief Executive Officer · 2026-05-07. In August he insisted “this pressure is not structural” — Chris Layden, Chief Executive Officer · 2026-08-06 and, in the same call, that “even in education, where we expect the second half of the year to flip back to growth” — Chris Layden, Chief Executive Officer · 2026-08-06. Three months later the pivot has become "more gradual," Q4 revenue guidance was revised, and the full-year margin outlook went to flat-to-plus-10 basis points. The structural-versus-cyclical question is now being answered by the calendar rather than by management.

We do view that as temporary. We're working, and they're still navigating a bit of a challenging market environment... it's a long-term client, a very large client and somebody we worked closely with for many years.

That quote is about the other guidance revision — an unexpected step-down at one large ETM client, which became the company's single most prominent keyword this quarter. Note the echo: a year ago, when three clients blew a hole in the numbers, the CFO told analysts “it's less about customer concentration, and it's more about stickier services” — Troy Anderson, Chief Financial Officer · 2025-11-06. Concentration keeps showing up anyway, which is why the market discounts the "temporary" label.

What Kelly stopped saying

The vocabulary changed this quarter, and that is its own signal. The growth office — the newly created commercial function that dominated the prior quarter's language, alongside a "commercial operating framework" and a unified CRM rollout — has slid out of the narrative almost entirely after one mention. In its place: a refreshed brand architecture collapsing every legacy name into "One Kelly," ten specialties, a new Kelly Digital Infrastructure unit, a Workday human-capital cutover, and AI tools named Grace Boost and Sevayo. Rebranding mid-recovery costs real money: $5.3 million of charges in the quarter and depreciation and amortisation up to $15.2 million from $12.6 million on accelerated trade-name amortisation.

The AI language, though, is where Kelly sounds least like itself and most like everyone else. Scan the reporters from the past five days and the words "AI agent," "agentic" and "personal agent" appear from Airbnb to Zebra. Kelly's twist is a trust pitch rather than a technology pitch.

Simply put, humans lead the process and AI elevates it. No matter how a candidate reaches us through a branch, an MSP program or our AI recruiting agent, every hiring decision is made by a member of our team.

That is a defensible position in staffing, where the product is judgment. It is not, however, a moat the tape is currently paying for.

The tape is long AI infrastructure and short Kelly's working capital

Kelly's stock is up 7.2% over the last 90 trading days into this print, which sounds like conviction until you note it is 18.7% below its August 28 high of $17.28 and 55.6% below its 2018 peak. At a $441 million market cap the company trades at 0.1x revenue — the market is pricing a cyclical, not a transformation.

There is a symmetry worth flagging. Kelly's best growth story is a second-derivative play on the AI build-out: a renewed and expanded global MSP programme with a semiconductor manufacturer of nearly 40 years, a relationship with a large IT infrastructure provider that now spans an AI innovation centre, and a new broader digital infrastructure brand. Yet the global tape is quietly repricing that exact theme — AI data centers shows up as the single largest cluster in the 30-day decliners, with 55 negative names and a negative 30% return. If AI capex digests, the ETM momentum that just turned positive loses its tailwind.

The balance sheet is the other half of the story. Kelly used $55 million of cash in the quarter for seasonality and working capital, ending with $122.5 million of debt against $22 million of cash and $256 million of available liquidity. That is a full reversal of the pandemic-era fortress: effective net cash of −$50 million versus +$254 million in 2020Q3. Working capital intensity explains why a return to growth eats cash — receivables at 118% of revenue, up 10.8 points year over year.

So the honest read: Kelly's cyclical recovery is real, narrow, and financed partly by its own balance sheet, while the segment that was supposed to be the reliable ballast — the market leader with industry leading fill rates and 70% of districts still in-sourcing, that white space — just pushed its inflection out again. Management's claim that “we're at the early stages of a recovery” — Chris Layden, Chief Executive Officer · 2026-11-05 is supported by ETM and SET and contradicted by Education. For a $441 million staffing company, being half-right about the cycle is the difference between a re-rating and another year of "temporary."