The seven numbers a restaurant group needs to see every week

The seven numbers

1. Revenue per location and platform

Not network revenue, but the matrix: twelve locations times three platforms gives 36 numbers, each with its deviation from the previous week and from the four-week average. A location that loses 20 percent on Uber Eats while eat.ch stays flat has a platform problem, not a location problem. You only see that distinction in the matrix.

2. Contribution margin per order

Order value minus commission, food cost, packaging, promotions and ads, divided by the number of orders. It is the most important number on the list, because revenue says nothing without it. How to calculate it is covered in our article what is really left of an order. Sample target: at least 28 percent of the order value before preparation.

3. Orders and average order value

Both numbers belong side by side. More orders with a falling order value often mean a promotion that attracts small baskets. Fewer orders with a rising order value can mean a minimum order set too high. Only the pair shows what happened.

4. Availability

Two parts: the hours the location was online on the platform compared with its opening hours, and the number of items that sat offline. A location that pauses for 40 minutes on Friday evening loses the best hour of the week, and no revenue report names that as the cause. Sample target: 98 percent online time, zero bestsellers offline.

5. Delivery time

The median time from order to handover, per location and platform. Not the average, because single outliers distort it. The apps weigh delivery time in their ranking, and guests rate it in stars. In a network of twelve locations the value fell from 45 to 32 minutes within three months, and revenue rose by 33 percent in the same period. Measured with one client, not a promise of the same result.

6. Rating

The average of the last 30 days per location and platform, plus the number of new ratings and the number left unanswered. The all-time average is slow and hides that one location has been sitting at 3.9 for three weeks. Every rating below four stars needs a reply within 48 hours and a cause.

7. Promotion and ad cost per order

Discounts and sponsored placements, added up and divided by orders. This number rises quietly. A location manager switches on a promotion, the platform suggests an ad package, and three months later marketing and discounts cost five francs per order. Sample target: at most 6 percent of the order value, every promotion with its own calculation.

Every number has a target and a name

A number without a target is information, not steering. Each of the seven numbers gets a target per location and a traffic light: green means carry on, red means someone acts. And each number has exactly one owner who can explain on Monday why it stands where it stands. Not the team, not head office, one person.

Targets differ by location. A location in the business district with lunch trade has a different order value from one in a residential area with evening trade. That is why we group locations by pattern and set targets per group.

The routine: 30 minutes on Monday

The review is prepared before the meeting starts. Nobody opens a dashboard in the meeting and searches. The sequence is the same every week:

1. 5 minutes. The red numbers, only those. What happened, who takes care of it, by when.
2. 10 minutes. Last week's promotions with proof: discount cost against additional orders. What does not pay gets stopped.
3. 10 minutes. One decision for the coming week: a price, a minimum order, an availability rule, not five.
4. 5 minutes. Open items from last week, done or not.

After 30 minutes there is a list with names and dates. Once a month a written report is added: what moved, why, what comes next, plus the comparison with the nearest competitors in the same segment.

Where the data comes from

The seven numbers need three sources: the platforms' partner portals (orders, order value, delivery time, ratings, availability), the till (food cost, preparation time, direct orders) and the accounts for commissions and invoices. At the start a weekly export into a spreadsheet is enough. Once the routine holds, a dashboard that merges the data automatically and sets the traffic lights pays off. The order matters: first the routine, then the tool. A dashboard without a Monday meeting is a screensaver.

Three mistakes we see in almost every network

• The average. Network figures hide the three locations that carry the result and the three that cost it.
• Revenue without margin. A 15 percent revenue increase from promotions can be a profit decrease. Without contribution margin per order nobody notices.
• Review without decision. An hour of looking at numbers and nothing on the list at the end. Every red number needs a name and a date.