Somebody asks how the weekend went and you answer, 'Busy.'
That might be true. But was it profitable? Which items made money? Did the Friday promotion bring in additional sales or just lower the price for customers who would have shown up anyway?
If your point-of-sale system can print a receipt, it probably records more information than you're using. You don't need to become an analyst to start asking better questions of it.
Sales are not profit
A burger that sells for $14 doesn't contribute $14 to overhead and profit. Food, packaging, payment costs, waste and sometimes extra labor all matter. Some costs vary with every order; others are fixed for the shift.
Before you judge a special, work out a rough contribution margin: selling price minus the costs that rise because you sold that item. Keep labor treatment consistent, and don't pretend a quick approximation is a complete P&L.
An item with strong sales but weak contribution can be useful as a draw. Just make that an intentional decision, not a pleasant surprise in the gross sales report.
Ask a question before opening a dashboard
Start with something practical: Which two hours on Wednesday justify the second bartender? How much of our food revenue comes from the five best-selling items? Are discounted drinks increasing total tickets? Do customers order food later when there's an event?
Then find the report that helps answer that question. Look at sales by hour, item mix, voids, discounts and payment fees where your system supports it. Match that against schedules and purchasing. A POS by itself doesn't know the whole story.
If your menu costs haven't been updated since last winter, don't treat calculated margins as truth. Bad inputs make very confident-looking bad reports.
Use a small scoreboard
Pick four numbers you can check weekly: gross sales, transaction count, average ticket and a sensible margin or labor measure. Add a specific event metric when you're running an experiment.
Suppose Tuesday sales rose 20 percent, but the number of transactions didn't move and discounts grew. That's a different story from a Tuesday when twenty first-time customers came in at normal pricing. The same topline can hide very different realities.
Also check for refunds, employee meals, split tickets and other quirks that affect comparisons. Make sure you're using the same time periods and definitions every week.
Don't collect data you won't act on
A hundred-chart dashboard is less useful than one report that changes a scheduling or menu decision. Each week ask: What surprised us? What might explain it? What are we changing? When will we check again?
And be careful with customer-level data. Collect only what you need, restrict access and comply with payment and privacy obligations. You do not need to identify every guest to learn when the kitchen is busiest.
Start with last month's ordinary Tuesdays
Export those shifts, compare the hours, staffing and item sales, and write down one hypothesis. Maybe late kitchen hours don't cover their variable costs. Maybe the first ninety minutes after work are stronger than you thought.
Test one change and see what happens.
Technology won't manage the bar for you. But it should help you make a decision with something better than 'felt busy.'
