Stop Guessing: The Data Behind Every Profitable F&B Decision
- David Holden
- Jul 17
- 5 min read
We were running a very busy steakhouse, selling around 900 fillet steaks per month. Every week, in the run-up to the weekend, I'd sit down with the chef before he placed the meat order, not to check up on him, but in order to ensure that we were ordering the right quantities of cuts to meet demand - Rump, sirloin, ribeye, fillet, chateaubriand. If we get that wrong we're either out of sirloin on a Saturday night with 40 covers still to come in, or sitting on a mountain of ribeyes because bookings came in lighter than last week.
That conversation between front of house and kitchen is forecasting. Not a spreadsheet that nobody in the kitchen has seen. A number, cross-referenced against the trend, agreed by the two people who actually have to live with it.
Most operators don't forecast like that. They look at what's on the books, make a rough call on how busy the week looks, and set their spend from there. It works until it doesn't - and when it doesn't, it knocks a hole in your bottom line.
Forecasting isn't a finance exercise.
It's the starting point for every spending decision you make before the week begins.
The £38 question
If your forecast assumes £38 spend per head and actual performance comes in at £35, that's not just a revenue shortfall. It means the assumptions underneath your forecast are wrong. Unless you find out why, every projection you build on top of it will be wrong too.
Was it the sales mix? Did guests trade down? Were fewer premium items sold? Did discounting creep up? These aren't just operational questions. They're forecasting questions, and they need answering the day the gap shows up - not at month end.
The data points that actually matter
Not every number in your system is worth tracking. The ones that drive a useful forecast are:
Spend per head — not as a weekly average, but by day type and session. A Saturday dinner spend per head looks nothing like a Tuesday lunch. Average them together and you've averaged out the insight.
Covers on the books versus your usual pick-up — how many additional covers typically arrive between your last count and service? That varies by day, season, whether there's a match on or a school holiday approaching. Without it, covers on the books is a starting point, not a forecast.
Walk-in percentage versus booked — some operations run almost entirely on bookings, others depend heavily on walk-in traffic. Knowing your split, and how it shifts across the week and the year, is fundamental to forecasting accurately.
Bar revenue — covers on the books tell you part of the story. Weather tells you the rest. A sunny Saturday with a terrace full of walk-ins drives a completely different bar number to the same covers on a wet Wednesday when everyone leaves straight after pudding.
External factors — weather, local events, road closures, a competitor opening nearby. None of it shows up in your booking system, all of it affects your covers and your spend.
The operators who log what was happening around them alongside their sales data build something invaluable over time. Not just a record of what they took, but a record of why. So when the sun comes out next August, they already know what happens to terrace covers, wet sales, and spend per head. They've seen it before. They planned for it. Everyone else is scrambling.
Same week last year is your most useful comparison
Not last week. Not last month. Same week last year.
That's the comparison that accounts for seasonality, for the rhythm of your specific trading pattern. If you're up 8% on the same week last year, apply that to your forecast and set your spend budget accordingly. If you're down, you need to know why before you commit to anything. Last week tells you what just happened. Same week last year tells you what to expect.
Your forecast holds the purse strings
This is the point most operators miss. A forecast isn't just a prediction of what's coming in. It's the basis on which you decide what goes out.
If you're forecasting £18,000 in revenue next week, your wage bill, order values, and prep quantities should all be derived from that number. Operators who don't forecast are writing blank cheques every week and hoping the revenue covers them.
The operators who get this right set a spend budget from the forecast before the week starts. Not after they've seen how it went. Before. That discipline is what separates the businesses that manage their margin from the ones that discover it at the end of the month.
What this looks like in practice
I maintained a forecasting spreadsheet across a multi-revenue-stream hotel operation covering breakfast, lunch, dinner, afternoon tea, bars, conferences and events.
Every day we tracked covers, spend per head, food revenue, liquor revenue and payroll by department against a target cost derived from the forecast. Kitchen, restaurant, bar and events were all measured separately.
If a department drifted over target, we saw it immediately and could act before it became a month-end problem. We also tracked third-party business separately from direct bookings, because commission impacts both revenue and margin. The same principle applied to events. Not all revenue delivers the same profit, and over time the data showed which business was truly valuable and which only looked good on paper.
The spreadsheet wasn't sophisticated software. It was consistent discipline. The value wasn't in the spreadsheet itself — it was in the history it created and the confidence that gave us when making decisions.
Patterns in your data are as close to certainty as you get
Gut feel is what you fall back on when you haven't got the data. It's not a strategy. It's a guess with experience attached. But if you've been tracking your numbers consistently and you've seen the same patterns repeat across two or three years, that's not a guess anymore. That's evidence.
You know what a sunny bank holiday weekend does to your terrace. You know what a wet Tuesday in January does to your wet sales. You know because you've seen it, logged it, and compared it.
That's the difference between forecasting and hoping. And it's why the operators who are close to their numbers consistently outperform the ones who are not.
Actuals vs forecast — every day
Compare your actuals against forecast every day, not at the end of the month.
Month-end figures are just the sum of every decision made during the period. They show where your KPIs landed, but if you're managing properly, you've seen the result coming for weeks.
By the time month-end arrives, the work is already done. The variances have been spotted, the conversations have happened, the corrective actions have been taken.
If month-end figures come as a surprise, your forecasting process failed somewhere in the previous 30 days.
The bottom line
You cannot manage what you cannot measure. And you cannot measure effectively unless you've been collecting the right data consistently over time.
None of this requires sophisticated software. It requires discipline, a consistent approach, and the commitment to stay close to your numbers every single day.
The operators who do that aren't guessing. They're working from evidence. That evidence improves forecasts. Better forecasts drive better decisions. Better decisions protect profit.
Data alone changes nothing. What matters is what you do with it.
The purpose of forecasting is not to predict the future. It is to give you enough warning to change it.
Comments