Analysis
Jun 30, 2026 · 8 min readHow to read restaurant same-store sales — and why most coverage gets it wrong
Same-store sales is the most-quoted number in restaurant earnings and the most misread. A comp is a traffic effect times a price-and-mix effect — decompose it into transactions and check before you judge the business. The analyst's lens, with worked examples from Starbucks, McDonald's, and Darden.
It is the single most-quoted number in restaurant earnings, and the most misread. A same-store sales figure is a traffic effect multiplied by a price effect — and the two routinely point in opposite directions. Here is how to read it like an analyst.
When a restaurant chain reports earnings, one number leads every headline: same-store sales (also called "comparable sales" or "comps") — the change in sales at locations open at least a year, stripping out the effect of opening new stores. It is the closest thing the industry has to a single health metric, and it is quoted everywhere. It is also, more often than not, read wrong.
The mistake is simple and pervasive: treating a comp number as if it means more customers. It frequently doesn't. A same-store sales figure is the product of two very different forces — how many transactions happened (traffic) and how much each one was worth (average check) — and a chain can post a positive comp while losing customers, simply by raising prices. Until you split the headline into those two pieces, you don't actually know whether the business is getting healthier or quietly hollowing out.
The formula that should sit under every comp number
Same-store sales growth decomposes cleanly:
Comp sales % ≈ change in transactions (traffic) + change in average check (ticket)
— and average check itself splits again into price (charging more for the same thing) and mix (customers trading up or down, or buying more or fewer items per visit). So a single comp percentage is really hiding three levers: traffic, price, and mix. The number that tells you about demand is traffic. The number that tells you about pricing power and inflation pass-through is check. A healthy comp is one carried by traffic; a fragile one is carried by price alone.
A same-store sales number is not one number — it is transactions (traffic) × average check (price × mix). Traffic is the demand signal; check is pricing power. The two routinely move in opposite directions, so a chain can post a positive comp while losing customers, simply by raising prices. Never read a comp as "more customers" until you've split it: a comp carried by traffic is real, a comp propped up by price alone is fragile. And never add same-store sales to total-revenue growth — comp excludes new units on purpose.
The decomposition that changes the story: Starbucks
Take a real example. In its fiscal first quarter of 2025, Starbucks (opens in new tab) reported North America comparable sales of −4%. Bad, but not alarming on its own. Now decompose it: that −4% was an 8% decline in transactions, only partly offset by a 4% increase in average ticket. Read correctly, the headline understated the problem — Starbucks wasn't down a little, it was losing customers at an 8% clip and using price to cushion the reported number. Over the following quarters the transaction decline narrowed (−3% in fiscal Q3, −1% by fiscal Q4, when comps reached flat) — and that improvement in traffic, not the ticket, was the real signal that the turnaround was working. The comp line alone would never have told you. The decomposition did.
Traffic is the truth-teller: McDonald's
The same lens, run over a year, explains McDonald's (opens in new tab) 2025. In the first quarter, U.S. comparable sales fell 3.6%, and the company was explicit that it was "primarily driven by negative comparable guest counts" — traffic, down about 2.6%. People were visiting less. By the fourth quarter, U.S. comps had swung to +6.8%, and McDonald's tied the recovery directly to value — the McValue platform and the relaunch of Extra Value Meals — which pulled guest counts back up. The lesson is the one the whole "value war" of 2024–26 turns on: traffic is the scoreboard, and value is the lever operators pull to move it. A comp propped up by price can flatter a chain that is actually shedding customers; a comp driven by traffic is real. Watch the guest-count line, not the headline.
Four more traps that mislead
1. Comp is not total revenue. Total sales growth includes new units; same-store sales deliberately excludes them. A chain can grow total revenue smartly by opening stores while its comps are flat or negative — the existing base is stagnating even as the footprint expands. Conflating the two makes a stalling brand look like a growing one. Always separate unit growth from comp.
2. The 53rd-week artifact. Restaurant fiscal calendars occasionally carry an extra week, which inflates reported totals with no underlying improvement. Darden's (opens in new tab) fiscal fourth quarter of 2026 showed total net sales up 13.7% — but part of that was an extra week in the calendar, not 13.7% of real growth. Same-store sales (which compare like periods) are the cleaner read; headline revenue growth in a 53-week year needs an asterisk.
3. Inconsistent definitions. "Comparable sales" is not standardized. Some chains report system-wide comps (all stores, including franchised) — the basis you'll see for franchisor-heavy brands like Yum! (opens in new tab) and Domino's (opens in new tab); others report company-operated comps only. The base ("stores open 13+ months") and the treatment of closures, remodels, and delivery vary by company. Comparing one chain's comp to another's without checking the definitions is comparing two differently-drawn lines.
4. Mix is not price. Within "average check," a rise can come from genuine pricing or from mix — customers buying premium items, or simply more items per order. A check up on mix (people buying more) is healthier than a check up purely on price (people paying more for the same), and a check propped up while items per transaction fall is a warning. The blended ticket hides which is happening.
Same company, two stories
One last discipline: a blended company number can bury divergent realities. Darden in fiscal 2026 grew Olive Garden same-store sales +4.0% for the year while LongHorn Steakhouse rose +7.2% — and in the fourth quarter the gap was starker, Olive Garden +2.4% (a miss) against LongHorn +9.5% (a beat). Same parent, same quarter, two different demand stories. Read the brand-level comps, not just the consolidated line.
The takeaway
A same-store sales number is a traffic effect times a price-and-mix effect, and the disciplined reader never accepts the headline without splitting it. Do decompose every comp into transactions and check before judging the business; do watch traffic as the real demand signal and treat a price-only comp as fragile; do strip new-unit growth and calendar artifacts from "total sales"; and do check the comp's definition before comparing two chains. Don't read a positive comp as "more customers," don't confuse same-store sales with total revenue growth, and don't compare comps across inconsistent bases. Run that discipline and the most-quoted, most-misread number in the industry turns into one of the most revealing — it stops being a headline and becomes a diagnosis.
This is the lens FoodBud applies to every earnings report we cover — because what a number means depends entirely on what's underneath it. See it at work in our deep-dives on Yum! (opens in new tab), Domino's (opens in new tab), and Starbucks (opens in new tab), and in the analysis Who really owns "China"? (opens in new tab)
The decomposition, worked (real examples)
| Chain (basis) | Headline comp | Traffic (transactions) | Check (ticket) | What it really says |
|---|---|---|---|---|
| Starbucks · NA, FQ1 2025 · company-operated comp | −4% | −8% | +4% | A traffic problem (−8%) masked by price; the headline understated it |
| Starbucks · NA, FQ4 2025 · company-operated comp | flat | −1% | +1% | Traffic decline nearly healed — the real turnaround signal |
| McDonald's · US, Q1 2025 · system-wide comp | −3.6% | negative guest counts (~−2.6%) | — | Comp fell because customers visited less |
| McDonald's · US, Q4 2025 · system-wide comp | +6.8% | guest-count growth (value-driven) | — | Value (McValue) pulled traffic back — a real recovery |
| Darden · Q4 FY2026, total net sales · 53-week | sales +13.7% | — | — | Inflated by a 53rd week — not 13.7% of real growth |
| Darden brands · FY2026 · company-operated same-restaurant sales | Olive Garden +4.0% · LongHorn +7.2% | — | — | One company, two divergent stories — read brand by brand |
Notes. Comp / same-store sales ≈ change in transactions (traffic) + change in average check (price × mix). Basis varies by company — system-wide (all stores, incl. franchised) vs company-operated only; comparable base ("open 13+ months"), and the treatment of closures/remodels/delivery, also differ — so compare like with like. "Total net sales" is not a comp (it includes new units and calendar effects). Figures from company releases / SEC filings (2025–FY2026).
Sources. Starbucks fiscal 2025 results (FQ1 NA comp −4% = −8% transactions + 4% ticket; FQ3 transactions −3% / ticket +1%; FQ4 comp flat, −1% transactions / +1% ticket, fourth consecutive quarter of improvement) — Starbucks investor releases / SEC 8-Ks. McDonald's 2025 results (US Q1 comp −3.6% on negative guest counts, traffic ~−2.6%; Q4 US comp +6.8% on value-driven guest-count growth) — McDonald's Q1 / Q4 2025 earnings releases / SEC. Darden fiscal 2026 Q4 + full year (total Q4 net sales +13.7% with an extra week; Olive Garden FY +4.0% / Q4 +2.4%; LongHorn FY +7.2% / Q4 +9.5%) — Darden June 25, 2026 release / SEC 8-K. Industry traffic −0.8% (2025) — NRN / QSR industry data. All USD; figures source-backed; comp definitions noted; no fabrication.