KELYA · Kelly Services, Inc. · Stock research

and# KELYA — Kelly Services, Inc.

Current Thesis

The post-print leg extended. Price closed 2026-08-21 at $17.09 — a new 52-week high, 0.0% below it — versus the $16.06 close of 2026-08-14 cited in the prior note, when the 52-week high stood at $16.31. Nothing company-specific arrived in that window: the last dated sponsorship action was Noble Financial's 2026-08-12 FY2026 EPS revision to $1.21 from $1.16, and the last target move was Barrington Research's 2026-08-10 raise to $20 from $15. The 2026-08-19 ex-dividend for the $0.075 quarterly payout came and went. What moved in that gap was price, running ahead of published estimates rather than alongside fresh ones.

The narrative is maturing, downgraded from the accelerating label carried on 2026-08-15. The evidence dating that change is the absence of new inputs after 2026-08-12 while the shares advanced to a high, plus a published mean analyst target of $15 with an average Hold rating (MarketBeat / WallStreetZen, August 2026) that now sits below the last close. The operational story still works; the flow of new headlines behind it has thinned.

Standing correction to prior coverage. The earlier note described KELYB as under Kelly-family control with "no activist path." That is no longer accurate. On 2026-01-30 the Terence E. Adderley Revocable Trust K closed a sale of 3,039,940 Class B shares to Hunt Equity Opportunities, LLC, making Hunt the controlling stockholder with 92.2% of the voting class. The board adopted a rights plan on 2026-01-12 while it evaluated the transaction; Amendment No. 1, effective 2026-01-30, exempted the Hunt purchase and provided for the plan to expire immediately prior to the transfer. Effective 2026-01-30 the board comprised four Hunt designees plus three incumbent directors. Class A holders own the economics of whatever Hunt decides and vote on none of it.

Bullish and bearish views on Kelly Services, Inc.

The model's bull view on Kelly Services, Inc. (KELYA), in brief: Q2 2026 adjusted EPS $0.37 against $0.24 consensus, on revenue of $1.041B versus $1.009B expected (2026-08-06 release). The bear view: H1 2026 adjusted EBITDA was $46.9M, down 34.8% YoY at a 2.3% margin — 90 bps of compression. Both cases follow in full.

Bull Case

• Q2 2026 adjusted EPS $0.37 against $0.24 consensus, on revenue of $1.041B versus $1.009B expected (2026-08-06 release). Revenue fell 5.8% YoY, an improvement of roughly 500 bps in the rate of decline versus Q1 2026.
• Q3 2026 guidance: total revenue $916.3M–$935.0M, underlying revenue growth of +1% to +2%, adjusted EBITDA margin in the low-2% range for a 40–50 bps YoY improvement (2026-08-06). First guided growth after roughly two years of decline.
• Temporary help services payrolls rose 3,400 in July 2026 and have now increased for seven consecutive months, +53,600 over that span, while total nonfarm payrolls fell 23,000 in July and the prior two months were revised down a combined 103,000 (BLS Employment Situation, released 2026-08-07). The leading staffing series is doing the opposite of the headline.
• Gross profit rate 20.4% in Q2, down 10 bps YoY but up 150 bps sequentially; adjusted EBITDA $31.1M at a 3.0% margin, also up 150 bps sequentially (2026-08-06).
• SET revenue declined 6.1% reported and 3.0% excluding discrete impacts but grew sequentially for the first time in two years; ETM grew 3.1% YoY excluding discrete impacts (Q2 2026 presentation).
• $279M cash, $43M drawn on US credit facilities, $303M total liquidity at quarter-end, with standby letters of credit at $78M versus $102M at year-end 2025 (2026-08-06). Recovery is not liquidity-constrained.

Bear Case

• H1 2026 adjusted EBITDA was $46.9M, down 34.8% YoY at a 2.3% margin — 90 bps of compression. The full-year guide of +10 to +20 bps requires H2 to more than reverse that, so the entire margin recovery is back-half loaded and unverifiable until ~November.
• The last close of $17.09 is above the mean published target of $15 (August 2026). Barrington's $20 is the single visible figure above spot, and no bulge-bracket house covers the name, so the estimate base that would justify the re-rate is two boutiques deep.
• RSI(14) at 76.7 with the shares up 61.1% over three months and sitting exactly at the 52-week high leaves no cushion inside the recent range for a macro disappointment.
• The Q3 revenue guide is not unambiguously above the Street: the $935.0M top exceeds one consensus figure reported 2026-08-06 ($929.3M) but sits below another cited the same day ($939.3M).
• Education revenue fell 4.4% in Q2 on what management called delayed contract decisions. The autumn school-year ramp determines whether that was timing or demand, and no interim datapoint resolves it.
• At a 3.0% adjusted EBITDA margin, a low-single-digit revenue miss swings EPS disproportionately; the estimate cleared in Q2 was $0.24, small enough that the percentage beat overstates the dollar move.
• Hunt's 92.2% voting control means any capital-structure decision, related-party arrangement or take-private terms are set without Class A consent.

Setup & Price Structure

The 2026-08-21 close of $17.09 is itself the 52-week high, which makes the structure young and untested — there is no overhead supply above and no confirmed shelf beneath. The reference level is the prior 52-week high of $16.31 that stood as of 2026-08-14; the breakout through it happened in the week ended 2026-08-21 without an accompanying company headline.

Crowding and positioning observables, stated as observables: RSI(14) 76.7; price 0.0% from the 52-week high; a three-month price change of +61.1%; the last analyst action dated 2026-08-12, eleven days before this writing; a mean published target ($15) below spot with the only above-spot figure ($20) dated 2026-08-10; boutique-only coverage; a dual-class float where the voting class is 92.2% held by one party since 2026-01-30 and therefore does not trade freely. No earnings date falls inside 30 days. The 2026-08-19 ex-dividend has passed.

Catalyst Calendar (next 30 days)

• 2026-08-28, 10:00 ET — BLS Current Employment Statistics preliminary benchmark revision (March 2026 reference), released alongside Q1 2026 QCEW data.
• 2026-09-04, 08:30 ET — BLS Employment Situation for August 2026. Temp-help services payrolls print here; the July series showed a seventh straight monthly gain.
• ~2026-09-16 (est.) — FOMC decision (meeting ~2026-09-15/16, est.). Sets the discount rate applied to a small-cap cyclical with no secular bid.
• No company-specific event. Next is the Q3 2026 print, ~2026-11-05 (est.), roughly eleven weeks out.

What Would Change Our Mind

The whole structure rests on a breakout above $16.31 executed in one week with no news attached to it. If that break fails, the case reverts to a deep-value staffing name with two boutique sponsors and no company catalyst for eleven weeks. A weekly close below $15 marks that failure: it returns price inside the range that held before the 2026-08-06 print and unwinds the August leg.

Three non-price conditions carry equal weight. First, the 2026-08-28 benchmark revision — a large downward revision to prior payroll levels would remove the labor-stabilization framing that supports the +1% to +2% underlying growth guide, and it lands in five days. Second, temp-help payrolls contracting in the 2026-09-04 report would break a seven-month sequence that is currently the strongest evidence for the guide. Third, at the Q3 print, reported revenue under the $916.3M floor or adjusted EBITDA margin below the guided low-2% range would show the back-half margin recovery did not happen.

The label flips to saturated if the shares hold above the mean published target with still no new initiation through the September quarter — mainstream pricing on unchanged sponsorship.

Correlation Notes

The name trades with the temp-help payroll series more tightly than with the broad tape; the 2026-09-04 report and the 2026-08-28 revision are shared catalysts for RHI, MAN, HSII and KFY, and a staffing-wide move on either date would say nothing about Kelly specifically. Small-cap value and the rate path are the second axis: no offsetting secular demand story insulates the multiple.

Two idiosyncratic wrinkles. KELYB is not a usable read-across — 92.2% has sat with one holder since 2026-01-30, so any A/B spread reflects float mechanics, not a governance signal. And thin turnover in the Class A means single-session moves can be outsized on modest volume in either direction, which cuts against reading any one day's tape as information.