The Capital Grille went 35 years without a happy hour. It launched one this year. When a steakhouse group that disciplined starts discounting weekday afternoons, that is not a trend piece, it is a market signal: the guests you want at 7pm are now deciding at 4pm based on price.
The numbers
Restaurateurs surveyed for OpenTable's 2026 Dining Trends report put happy hour and value promotions as the number one trend for 2026. The same data shows dining between 4:00 and 4:59pm up 13% year over year. Per Nation's Restaurant News, nearly 30% of restaurant visits are now tied to a deal, the highest share in 50 years.
Does it work for the bar? POS provider Cake compared 400 bars with and without happy hour programs. The happy hour bars ran 33% higher transaction counts overall, and during the window itself they beat the others by more than 20% in revenue, even though the non-happy-hour bars were bigger earners the rest of the day.
So the traffic is real. The margin problem is real too, and it comes from how most bars implement it.
The wrong way: discount the list
The common version is "$5 off cocktails until 6." Take a $16 cocktail with $2.40 in liquid cost. That is 15% pour cost and $13.60 of contribution. Sell it at $9 and the same build runs 27% pour cost and contributes $6.60. You gave up half the margin, and you taught your regulars that the list price is negotiable. Guests who learn a drink is worth $9 on Tuesday resist paying $16 for it on Friday.
A discount on the core list is a price cut on your best products. That is the surrender version.
The right way: a separate program
- Write a separate menu. Five or six drinks that exist only in the window. The core list never changes price, so its price integrity survives. The happy hour card is a different product at a different price, not the same product marked down.
- Engineer builds to the price. Target the same pour cost discipline you run at night. At $9, that means $1.20 to $1.50 in the glass: batched stirred builds, house cordials, sensible base spirits, carbonated builds from the tap or bottle. Designed at $1.35, a $9 drink runs 15% pour cost. Same math as the $16 drink, smaller check, margin intact.
- Fence it hard. Weekdays, 3 to 6, bar and lounge seats only. The Capital Grille's version runs exactly that way: lounges only, weekdays 3 to 6pm, $10 cocktails and $8 wines on their own card. Fences are what keep the deal from leaking into prime time.
- Ladder, do not lower. The happy hour card is the bottom rung and the full list stays on the table. A guest who came for the $9 drink orders the $16 one when the second round lands after 6. Their bar menu at $9 to $19 sits under the dining room's reserve pours for the same reason: every price has somewhere to climb.
- Batch everything on the card. The window runs on your thinnest staffing of the day. Every build should pour in under a minute. If a happy hour drink needs a fresh citrus shake à la minute, it is on the wrong menu.
The window math
Core list drink: $16, cost $2.40, margin $13.60, 15% pour cost
Discounted to $9: margin drops to $6.60, pour cost 27%
Engineered $9 build: cost $1.35, margin $7.65, 15% pour cost
30 engineered serves in the window: $229 margin from dead hours
Fixed costs from 3 to 6pm are already paid. Every seat filled is found money.
Measure the window, not the feeling
Two numbers tell you if it works. Transactions between 3 and 6pm against your eight-week baseline: that is the traffic the program bought. And revenue between 6 and 8pm over the same comparison: if it holds or rises, the window is feeding prime time; if it drops, your fences are leaking and full-price guests are shifting into the deal.
Give it eight weeks before judging. Happy hour traffic is habit traffic, and habits take a few cycles to move.
Building the card, the batch specs, and the pricing tiers is menu work: Menu & Program Refresh. If your 3 to 6pm is dead and you are not sure a discount window is the answer, fifteen minutes will tell you.