CAVA's First Misstep: 1.9% Comps, an EPS Miss, and a 54% YTD Slide

I read the CAVA Q3 earnings release on the train back from a tasting in Tribeca, and the line that stopped me wasn’t the headline — it was traffic. Negative 1.5%. For a brand that built its whole narrative on Chipotle-style throughput and a Mediterranean cultural moment, a quarter where fewer humans walked through the door is the first crack in the story. The comp came in at 1.9% against a Street whisper closer to 2.8%, and the EPS line missed too. The stock, already down 54% year-to-date, didn’t need much of an excuse.

So here’s the contrarian thesis I want to stake out, because the bulls and bears are both going to overreact: CAVA’s Q3 is the first genuine misstep since the IPO, but it is a reset of expectations, not a structural break. Restaurant-level margin still printed 24.6%. New units are still opening on plan. Adjusted EBITDA grew 19.6% to $40M. The full-year guide is intact. What changed is the multiple the market is willing to pay for “fast-casual darling” — and that change was overdue, CEO transition or not.

The Number That Actually Matters

Let’s be honest about what spooked the tape. Revenue of $289.8M, up 20%, is fine. Seventeen net new units bringing the system to 415 is on plan. The full-year guide of 68–70 net new restaurants and $148–152M of adjusted EBITDA was reaffirmed. None of that is the problem.

The problem is the composition of the 1.9% comp. Traffic was -1.5%. Price and mix did all the lifting. For a category that spent 2023 and 2024 telling investors that Mediterranean was the growth lane in fast-casual — that CAVA could be a 1,000-unit brand because the cultural pull was real — a quarter where you have to lean on check to make the comp positive is a different story than the one the deck was selling. CNBC’s same-day write-up framed it as the first miss since IPO, and that framing is going to stick.

Margin is the bull’s reply, and it’s a good one. 24.6% restaurant-level margin in a quarter where labor and commodities are still annoying is genuinely impressive. It says the four-wall economics that justified the unit-growth thesis haven’t moved. If you believed CAVA at 600 units and 24% margins, the Q3 print didn’t break that math.

But — and this is where I land — the multiple was never priced on margins. It was priced on the assumption that traffic would compound mid-single-digits for years because the brand was still being discovered. One quarter of -1.5% traffic doesn’t kill that thesis, but it forces every sell-side model to add a “what if discovery is mostly done” sensitivity. And once that sensitivity goes in the spreadsheet, the multiple compresses regardless of what the next two quarters do.

Schulman’s Inheritance

The other reason this print matters more than a one-quarter miss usually would: CEO Brett Schulman is transitioning out, and Q3 is the last clean read on the Schulman-era operating system before someone else owns the P&L. (Mark the interpretation here — this is my read, not the company’s framing.) The retained guide is, in effect, a promise the incoming leadership now has to deliver against. The 54% YTD stock decline is the market pricing in the risk that the new operator will use the transition to rebase expectations lower, the way new CEOs almost always do in their first 90 days.

If I’m advising the incoming team, I take the rebase. Q4 is the cleanest possible kitchen sink. Pull forward any margin investments. Reset the unit-growth algorithm to “67 plus or minus” so 2026 has room. The cost of looking conservative for two quarters is small; the cost of missing again in Q1 with the new CEO’s name on the press release is enormous.

What I’m Watching Next

Three things, in order. First, the Q4 comp split — if traffic stays negative and price keeps doing the work, the bear case (CAVA is a regional brand with a coastal ceiling) gets oxygen. Second, the average-unit-volume disclosure in the 10-K. AUVs north of $2.9M kept the unit-growth math honest; any softening there matters more than the comp. Third, throughput tech. Sweetgreen’s Infinite Kitchen rollout is the obvious comparison point, and I’ll have a forthcoming May case study on what that automation actually did to four-wall margins — because if CAVA’s answer to slowing traffic is “more units faster,” the operational backbone has to scale, and right now that backbone is still mostly humans.

The next-day Motley Fool transcript will get parsed for every word Schulman said about traffic drivers. Listen for whether management blames weather, calendar, or category. Weather is a tell that they don’t have a thesis yet. Category is a tell that they do, and it’s worse than the stock currently reflects.

For now: first misstep, not first stumble. But the next print is the one that decides which.

— Luca covers restaurants for TableTransfers. Tips: [email protected].