Sweetgreen (SG) Q3 2026 earnings review
Sweetgreen's customers returned in October, but its costs keep climbing
Sweetgreen's customers started coming back after a summer food-safety scare, but its restaurants still make less money. Sales at restaurants open at least a year fell 6.4%, a smaller drop than a year earlier. Chief executive Jonathan Neman said October sales turned positive. Restaurant-level profit slipped to 10.4% of sales as labor and delivery costs climbed.
| Sales at restaurants open at least a year | (6.4)% (9.5)% a year ago |
|---|---|
| Total sales | $176.0 million +2.1% from a year ago |
| Share of sales kept after restaurant costs | 10.4% 13.1% a year ago |
| Full-year outlook | Raised established-restaurant sales now about (5.5)%, from (8.0)% to (7.0)% |
⚖️ Verdict: 🟢 Bullish
The standing case got better, thanks to the outlook rather than the quarter. Sweetgreen now plans for full-year sales at established restaurants of about minus 5.5%, well above the range it set three months ago. It raised its profit outlook too. One caution: it trimmed this year's opening plan again, so new restaurants will contribute less.
The question now is whether October's customers stay without a cheaper basket. One view: better food and service brought them back, so sales per restaurant can rebuild. The other: discounted wraps bought the visits, so each order shrinks. February's fourth-quarter numbers will settle it.
🐂 Bull Case
October Sales Turned Positive After the Summer Scare
In July, United States authorities identified a multistate outbreak of cyclosporiasis, a gastrointestinal illness. No Sweetgreen ingredient turned up in the investigation, the filing says, yet customers still backed away from fresh prepared food. Finance chief Jamie McConnell put the cost at about five percentage points of traffic.
"We believe we have moved past the sales impact of the Cyclospora headlines, and sales momentum continues to build," chief executive Jonathan Neman said. Sales at restaurants open at least a year fell 6.4%, and Neman reported that "October comparable sales were about 5% with 4% transaction growth".
The outlook now carries that claim. Sweetgreen lifted its full-year target for established restaurants to about minus 5.5%, two points above the middle of the range it cut three months ago. Neman said the plan "is beginning to take hold, with sales improving through the quarter and positive comparable sales in October".
What to watch: how the fourth quarter splits between visits, prices and product mix. McConnell plans for visits to add about 4 points and prices 3. By our math mix then subtracts roughly three points, which would mean returning customers are choosing cheaper items.
New Menu and Marketing Brought Back Lapsed Guests
Sweetgreen spent the quarter giving people reasons to walk in. It ran a collaboration with the cook Molly Baz, its first Chef in Residence, and kept pushing wraps, the cheaper hand-held format launched nationally in May.
Neman said the gains came mostly from existing guests ordering more often, plus "a huge step-up in lapsed guests", and he called Gen Z one of the fastest-growing groups. Brand surveys help explain it: "complaints around both portion size and pricing are now at their lowest level in three years".
What to watch: the Plates relaunch, the warm dinner platters Sweetgreen first sold in 2023, now led by a bigger rotisserie-style chicken. Sales after 5 p.m. are only about a third of the business, so a lift there would prove the brand can win dinner.
Peak-Hour Records at Forty Restaurants
Throughput, meaning how many customers a restaurant serves in its busiest hour, is the gauge Neman uses to judge whether a team is ready. Sweetgreen standardised preparation, staffing and peak readiness under a programme called Project One Best Way.
Volumes ran seasonally lower in September and October, yet "40 restaurants set new peak hour records over the last two months", Neman said. He argues that closing the distance to his best restaurants could add several percentage points to sales over time, while admitting the gap to the median store stays wide.
What to watch: progress toward promoting 90% of head coaches, the managers who run each restaurant, from inside. Neman set that goal for the coming year, and more internal promotions would show that training, not pay alone, is what steadies a restaurant team.
🐻 Bear Case
Average Sales Per Restaurant Fell Another 11%
Average unit volume is what a typical restaurant rings up over four quarters. Sweetgreen's figure is now $2.47 million, against $2.77 million a year ago, and it has fallen in each of the past six quarters.
Sales at restaurants open at least a year have been negative for seven quarters running, and what customers buy explains part of it. Wraps cost less than salads, and Neman said they "have created a mix headwind for us because they're lower price".
- Customer visits: down 4.9 percentage points
- Product mix: down 2.1 percentage points
- Average check: "our check is only up about 1%", McConnell said, after a 3% price rise in September
- Permanent closures this year: five, two of them this quarter
This matters because rent, equipment and a minimum crew barely move with volume. A typical restaurant now takes in roughly $300,000 less a year, and almost all of that drops out of profit. Neman frames the quarter as momentum building, yet the measure of how much business each restaurant does kept sliding.
What to watch: average unit volume in the full-year report. A figure that stops falling would show October's extra visits are adding sales rather than shifting them to cheaper items.
Labor and Delivery Costs Ate the Food Savings
Restaurant-level profit is what a store keeps after food, labor, rent and the other costs of running it. Sweetgreen kept 10.4% of sales, down from 13.1% a year ago. That fall of 2.7 percentage points came despite a genuine win on ingredients.
- Food, beverage and packaging: 28.8% of sales, 1.8 points better than a year ago
- Labor: 30.5% of sales, 1.4 points worse
- Other restaurant costs, mostly repairs and delivery fees: 19.9%, 2.4 points worse
- Rent and occupancy: 10.3%, 0.7 points worse
The filing attributes the labor line to "deleverage from lower sales volume, labor inefficiencies, and wage inflation", so two of those three causes sit inside Sweetgreen's control. Repairs and delivery fees did the rest. Because every cost line except ingredients moved the wrong way, the food saving bought nothing.
What to watch: labor as a share of sales in the fourth quarter, which McConnell plans at about 31%. He has also promised about 1.5 percentage points of savings on cost of sales during 2027; a labor line that keeps climbing would swallow them before they reach profit.
Sweetgreen Will Open Half as Many Restaurants
New restaurants are the only reason Sweetgreen's reported revenue still grows. They contributed $13.8 million this quarter, against a $10.7 million fall at restaurants open a year or more. Three months ago that contribution was $18.4 million.
Sweetgreen cut the opening plan again, to about 10 net new restaurants this year from roughly 13. McConnell said next year's net additions will run at "about half of the 2026 level". The filing lists three sites still under construction, against 11 at the end of December.
There is a defensible reason, because capital spending has come down by more than $40 million so far this year and Neman wants better returns before building again. But the cut also removes the growth that has been masking negative sales at established restaurants. Neman expects to step building up the year after next and reach a 10% unit growth rate by 2029.
What to watch: the number of sites under construction in the full-year report. If it stays near three, the new-restaurant contribution keeps shrinking through 2027 and reported revenue starts tracking established-restaurant sales alone.
Risks this quarter didn't answer
Four standing worries got no clear answer this quarter. The release and the filing detail costs and sales, but leave the cash position, the Wonder shareholding, the planning record and the customer detail where they stood.
- Cash burn: operating activities used $28.5 million over three quarters, against $4.0 million a year earlier; a fourth quarter that generates cash would settle it.
- Wonder shareholding: the $86.4 million of preferred stock taken for Spyce sits at cost with no adjustment; a priced financing Sweetgreen accepts as observable would test it.
- Planning record: the 2026 plan rose for the first time after a cut, and the fourth-quarter target needs a ten-point swing to prove it.
- Customer detail: Neman calls Gen Z a fast-growing group, yet Sweetgreen reports no sales by age, region or city.
👓 Other Themes
New Pay Plan for Restaurant Managers
Last week Sweetgreen introduced a pay programme that "transforms the earning potential of our head coaches", the managers who run each restaurant. Those with three or more years at the highest-volume stores can earn over $200,000. McConnell expects it to cost about 0.25 percentage points of restaurant-level margin next year, after "a special bonus incentive for our head coaches and an inventory adjustment" this quarter.
Tariffs No Longer Appear as a Separate Cost
Sweetgreen says tariffs had minimal net impact on the cost of building a restaurant this year, helped by advance purchasing and sourcing changes. On food and packaging, the filing says tariff costs "have been absorbed into our supplier pricing and are no longer separately identifiable", so investors can no longer track a line that hurt profits a year ago.
💲 Other KPIs
Sweetgreen's own app and website, plus loyalty scans at the counter, brought 38.0% of sales against 35.3% a year ago. The share has held near 38% for three quarters after climbing through 2025. Counting third-party delivery apps as well, digital reached 66.1%.
Head-office costs fell 10.5% and now take 15.7% of sales. The line covers operations, technology, finance, legal and brand marketing. Lower share-based pay drove $2.5 million of the drop, so most of the saving never touched cash. Underlying support-centre costs held near $24 million, roughly flat against last year.
Money owed to suppliers fell 39% from the year-end figure. Paying bills sooner drains cash, and the line took $5.8 million out of operating cash flow over the first three quarters. A year earlier it added $0.9 million.
🔮 Guidance
Raised. The plan now has sales at established restaurants down about 5.5% for the year, two points better than the range set three months ago. By our math, that leaves the fourth quarter near +4%, matching what McConnell described. It is the first increase to this year's plan, which Sweetgreen had cut three months earlier.
Raised. The range moved up half a point. The first three quarters came in at 11.3%, and McConnell put the fourth quarter near 11%. He said fourth quarters normally run about three percentage points below the year's average, so hitting the top of the range needs the cost savings to arrive early.
Raised. Earnings before interest, tax, depreciation and one-offs improve by $3.5 million at the middle of the range. The first three quarters lost $15.0 million, so by our math the fourth quarter loses about $6.5 million. A year ago it lost $13.3 million. Sweetgreen had cut this plan once, by $28.5 million at the middle.
Cut. Sweetgreen will add about 10 restaurants net this year, down from roughly 13. Nine arrived in the first three quarters. The filing lists three sites under construction, two of them due to open before the year ends. That is the third reduction to this plan in a year.
New. McConnell said next year's net additions will run at about half this year's pace, while Neman put gross openings near 10 with a few lease endings on top. Neman expects to step building up the year after and reach a 10% unit growth rate by 2029. Fewer openings mean reported revenue converges on what established restaurants do.
New. Three months ago the finance chief described this period only in words; now it carries a number. McConnell plans for customer visits to add about 4 points and prices about 3. Lapping a one-off gift-card benefit costs 1.5 points. By our math product mix then subtracts roughly three points, wider than this quarter's drag.
New. McConnell expects restaurants to keep about 11% of sales, up from 10.4% in this quarter, even though fourth quarters are seasonally the weakest. About 0.6 points of one-off costs sit inside the figure: a special bonus for restaurant managers and an inventory adjustment. Stripping those out, he put the underlying annual rate near 14%.
❓ Key Questions
Why does the fourth-quarter plan need a three-point mix drag?
Visits and sales are each planned at about 4% with prices adding 3%, so product mix must subtract roughly three points, worse than this quarter. Which items are pulling the average order down that hard?
When does create-your-own pricing reach every restaurant?
The commitment three months ago was a system-wide rollout by year end. Neman now says the first quarter, after 90 days in about 70 stores. What specifically still has to be fixed?
What cash return does a new restaurant earn today?
The 2025 class opened into a falling comparable base and the average restaurant now does $2.47 million. Sweetgreen has never published per-unit productivity or cash-on-cash returns for that class.
What would make the Wonder stake worth less?
The $86.4 million of preferred stock sits at cost with no adjustment and amounts to 72% of cash. What liquidation preference does it carry, and what would trigger an impairment review?
What was last year's gift-card benefit?
McConnell says lapping it costs 1.5 points of fourth-quarter comparable sales. The item was never broken out when it helped, so investors cannot judge how large the distortion was.