Restaurants & Food · Scenario 04

Izakaya / Bar

Drinks flow but the kitchen barely profits, and the average check has been flat for years. We mine what guests actually order together and rebuild the menu to lift every tab.

Method · Association-rule mining

The situation

An izakaya lives on the second and third round and the small plates that go with them. But the menu is a long, flat list, staff upsell inconsistently, and the combinations that naturally belong together are buried. Tables under-order not because they’re full, but because nothing nudges the next plate.

The owner suspects certain dishes ‘sell the drinks’ but has no evidence for which, so the menu and the specials board are designed by taste, not by what the receipts already reveal.

$34
avg check, flat 3 yrs
480
tabs / week analysed
56
menu items, unranked
Ad hoc
upselling by staff

Where we dig for the truth

We run market-basket analysis on every tab to find which items genuinely pull others along — the pairs and bundles worth building the menu around.

Itemised tabs (per table)Order sequence & timingDrink-to-food pairingsServer / section IDsDay & party sizeMargin per item
Which orders pull the next oneAssociation lift between a drink and the plate that follows0x1x2x3x4x3.1xHighball→Karaage2.6xSake→Sashimi2.2xBeer→Edamame2xYuzu sour→Gyoza1.8xShochu→Motsu

Highball drinkers order karaage at roughly three times the base rate. That pairing should headline the menu and be the server’s first suggestion.

Our approach — Market-Basket Analysis

Association-rule mining ranks every pairing by ‘lift’ — how much more likely B is once A is ordered. The strongest, highest-margin pairs become designed bundles, menu adjacencies, and a short, specific upsell line for staff (‘with the highball, the karaage?’).

We redesign the menu layout and specials board around these relationships, and set two or three courses that package high-lift, high-margin items at a price that still feels generous.

From a flat menu to a designed one1Itemise the tabsBreak every check intoitems and the orderthey arrived.2Mine the pairsCompute support,confidence and liftacross thousands ofcombinations.3Design bundlesTurn the besthigh-margin pairs intoset courses andadjacencies.4Script the upsellGive staff twospecific, data-backedsuggestions, not"anything else?".
Average check per tabEight weeks after the redesign$0$14$27$41$54W1W2W3W4W5W6W7W8S&I starts

The check climbs not by pushing harder, but by suggesting the very thing the data says the guest already wants.

What changes

Same menu items, re-arranged and bundled around real ordering behaviour. Representative for a 40-seat izakaya.

Representative 90-day movementAvg. check$34$47▲ +38%Plates / table4.15.3▲ +1.2Kitchen margin54%63%▲ +9 ptsMonthly profit$16k$27k▲ +69%
Where the uplift comes from+$11kmonthly profitDesigned bundles40%Menu adjacency30%Staff upsell script30%
Why this is not "social media management"
We didn't run a happy-hour gimmick or buy followers. We read the receipts the bar already had, found which plate sells the next drink, and rebuilt the menu around the evidence. That's analytics doing a marketer's job.

Frequently asked questions

How do you increase the average check at an izakaya?
We run market-basket analysis on thousands of tabs to find which drinks and plates are genuinely ordered together, then build designed bundles, menu adjacencies and a short data-backed upsell script around the highest-margin pairs.
What is market-basket analysis?
It is a data-mining technique that measures 'lift' — how much more likely one item is to be ordered once another has been. It reveals the pairings worth building a menu around, such as a highball and karaage.
Is this just running a promotion?
No. There is no discount or ad spend. We read the receipts the bar already has and redesign the menu around proven ordering behaviour — analytics doing a marketer's job. Book a marketing audit.

Want this run on your numbers?

Give us a month of itemised tabs and we’ll show you the pairings worth a redesign.

Book a marketing audit
Back to Restaurants & Food