How the best operators manage by the numbers.
The KPIs that actually move a shift — and the ones that just fill a dashboard. What shift managers, store managers and ops directors should be watching, and when.
Most restaurants have too many numbers and not enough signal. A busy operator doesn't need a BI degree — they need nine numbers, checked at the right moment, with a clear idea of what to do when one of them drifts.
Everything else is noise.
After walking the floor of hundreds of restaurants, the same nine numbers come up again and again. Not because they're clever — because they're the ones that change a decision in the next hour.
Sales vs last week
Week-on-week is the most honest comparator — same weather, same trading pattern, same offers. If you're down, there's a reason. Find it today, not at the end of the month.
Sales vs last year
Year-on-year removes the noise of a quiet week and shows the real trend. Good operators know their YoY to the pound before Monday morning.
Tickets (order count)
Sales without ticket count is half a story. If revenue's up but tickets are flat, something's moved on AOV. If tickets are down but sales are flat, you're over-reliant on big baskets.
Average order value
AOV tells you if the upsell is working, if the menu engineering is right, and if aggregator mix is dragging you down. Watch the trend, not the single number.
Door-to-door time
Total time from order received to customer handover. Under 40 minutes is the benchmark. Over 45 and repeat rate falls off a cliff.
% under the target
The distribution matters more than the average. "Average 38 minutes" hides the 20% of orders that went out at 55. That 20% is your churn.
Make, rack & drive time
The three components of door-to-door — kitchen prep, holding on the pass, driver route. Each one has a different fix. Without the split, you're guessing.
Labour cost %
The second biggest controllable line after food. Good operators know their labour % by shift — not just by period — and move people before the wage runs away.
Food cost %
The biggest controllable. Ideal vs actual is the number that quietly kills margin. A 2% slip on food cost is a 2% slip on profit — there's no margin to absorb it.
Three layers. Three cadences.
The same numbers matter to everyone. But when you look at them, and what you do next, changes with your role. This is the rhythm good operators run on.
The shift manager — minute by minute
Watches: live order queue, door-to-door time on the next five tickets, kitchen make time, driver availability.
Acts: re-routes drivers, pulls a second person to the pass, pauses aggregators if the kitchen is drowning. The goal is to get the next ticket out on target — not to fix the week.
The store manager — daily
Watches: yesterday's sales vs last week, % under 40 minutes, labour cost for the shift just closed, food waste, refunds.
Acts: adjusts the rota for tomorrow, has the one-to-one with the kitchen lead, signs off on the food order. Protects tomorrow, not yesterday.
The ops director — weekly
Watches: every store's WoW and YoY, aggregator mix, repeat rate, P50 / P90 delivery times across the estate, food cost variance.
Acts: allocates marketing spend, spots the outlier store, decides which playbooks travel and which stay local. Protects the quarter, not the week.
Check the right number at the right moment.
If a shift manager is checking labour cost every hour, they're doing the store manager's job. If a store manager is checking the next ticket's door-to-door time, they're doing the shift manager's job. The skill is knowing which number is yours, right now.
The best operators have three screens in their head:
- The next hour — what's on the pass, who's on the road, which kitchen station is slow.
- The next shift — who's in, what the food order looks like, where yesterday bled margin.
- The next week — where the estate is trending, which stores need attention, where marketing spend gets the best return.
That's it. Everything else is a dashboard you'll stop looking at by week three.
The traps that steal attention.
Averaging everything
The average hides the outliers. "Average delivery time 38 minutes" sounds fine — until you see the long tail at 55+ that's driving your churn.
Looking at sales, not tickets
Sales up, tickets down is a fragile number. One good day from a big group order can mask a slow week of real demand.
Ignoring the mix
If aggregator share is creeping up and own-brand is flat, your margin is quietly being eaten — even if top-line looks healthy.
No split on door-to-door
A 45-minute delivery can be a slow kitchen, a slow pass, or a slow driver. Three different fixes. Without the split you guess.
Monthly numbers, weekly decisions
If you only see the number at month-end, the bleeding already happened. The best operators close every week on Monday morning, not 20th of the next month.
Dashboards no-one uses
A dashboard you don't open isn't a management tool, it's decoration. One number on one screen beats forty numbers on ten.
If you don't know these, find out.
- What % of my orders last week went out in under 40 minutes? (Not the average — the percentage.)
- What's my WoW sales figure for the last four weeks, and do I know why?
- What's my labour cost % and food cost % for last week, and who owns each?
If those three answers aren't on your phone in 30 seconds, your tech stack is in the way. That's not a feature problem — it's a foundation problem.
More from Table Talk.
Do you need more orders? Think again.
Chasing new customers is expensive. Keeping them is where the money is. The maths behind why the best operators obsess over retention, not acquisition.
Back to Table Talk.
Every post in one place — growth, numbers, service, margin, tech. Notes from the people who've run restaurants.
Want to see these nine numbers on one screen?
Book a 20-minute demo. We'll show you how Andromeda surfaces the KPIs that actually move a shift — for the shift manager, store manager and ops director, each at the right cadence.
Book a demo