- Market
- Saudi Arabia
- Written for
- COO, Commercial Director
- Decision
- See what actually sells, when, where and at what contribution
- Languages
- EN · AR
The sources
Five feeds that must agree before a dashboard deserves trust
Every number the operator will act on converges from systems that were never designed to agree. The analytics work is mostly here — reconciling the feeds — and only lastly in the charts.
Five inputs — POS transactions, direct orders, aggregator statements, recipes and stock, rosters — converge into one reconciled model, which feeds menu decisions, channel decisions and branch actions.
POS transactions
Item-level sales, voids, discounts, timestamps — the ground truth of what left the kitchen.
Direct channel orders
Web and app orders with the customer attached, joining sales to identity.
Aggregator statements
Order exports and settlement files from the delivery aggregators, each in its own shape.
Recipes and stock counts
What each dish should consume, against what the store room says was consumed.
Rosters and trading hours
The denominator — branches compared without their hours and staffing are compared unfairly.
What we do
One reconciled model
Items mapped across channels, dates and branches aligned, settlements tied back to orders — the unglamorous layer every honest chart stands on.
Menu decisions
What to reprice, promote, rework or retire — argued from contribution.
Channel decisions
Which demand each channel should carry, seen in its own economics.
Branch actions
The specific branch and daypart where a number moved, not a chain-wide average.
Menu engineering
The bestseller that loses money every time it sells
A sales report ranks dishes by how often they sell. Menu engineering asks a colder question: after ingredients, packaging and the channel it sold through, what did each plate actually contribute? The two rankings rarely match. A generously portioned signature dish can dominate volume while contributing less per plate than the side order beside it, and a quiet item nobody promotes can be quietly carrying the margin. Plotting contribution against volume sorts a menu into its real four corners — the stars to protect, the workhorses to nudge on price, the puzzles to promote, and the losses being subsidised out of habit.
The same dish also earns differently by daypart and by branch. A breakfast item at noon, a dine-in dish sold through a delivery aggregator with its packaging and commission attached, a branch whose mix skews to the channel that pays least — averages hide all of this, which is why the layer we build always answers where and when, not just what. The comparisons stay honest because the baselines do: branches are read against their own hours, their own channel mix and their own history before they are read against each other.
Waste closes the loop. Recipes say what a dish should consume; stock counts say what the store room actually lost. The gap between them, tracked per branch and per week, is a signal worth more than most dashboards — sometimes it is over-portioning, sometimes prep planned for a Friday that did not come, sometimes a colder problem. The analytics do not name the cause; they name the branch, the week and the size of the gap, which is exactly what an operations lead needs to go and look.
The daily habit
Four views worth opening before the doors do
A dashboard that is not opened daily is a report. These are the views we design with the operations lead, cut to the decisions a morning actually contains.
Yesterday, reconciled
Sales by branch and channel with voids, refunds and aggregator adjustments already netted — the number that will survive month-end, not the flattering gross.
The daypart clock
How each daypart traded against its own baseline, so a soft lunch shows up as a soft lunch rather than vanishing inside a good dinner.
Channel economics, side by side
Direct and aggregator demand per branch, each carrying its own costs, so the mix is a number the operator watches rather than a feeling that surfaces at renewal time.
The exceptions worth a phone call
The waste gap that widened, the item voided unusually often, the branch whose refunds spiked — surfaced by rule, so the lead reads five exceptions instead of fifty charts.
Our role
Analytics we wire in, decisions the operator keeps
- 01
The reconciliation layer
We map items across channels, align branches and dates, and tie settlements back to orders — the work that decides whether every chart above it can be trusted.
- 02
Views cut to decisions
We design the dashboards with the people who will open them, and remove what no decision reads — a smaller dashboard that is used beats a complete one that is admired.
- 03
The operator's own literacy
We hand over the model, the definitions and the training so your team extends the questions without us — analytics that only a vendor can query is a subscription, not a capability.
What to reprice, which channel to steer and when to intervene in a branch are operating judgements; we make sure the picture those judgements are made from is complete, reconciled and opened every day.
Common questions
What buyers ask
Our POS already produces reports. What does an analytics layer add?+
The POS reports on itself, honestly and narrowly. It cannot see the aggregator settlements that arrive in their own files, the direct orders with a customer attached, the recipes and stock counts that turn sales into contribution, or the rosters that make branch comparison fair. The analytics layer exists to reconcile those sources into one model — item names mapped across channels, adjustments netted, baselines set — and then to ask the questions no single system can: what did this dish contribute, in this daypart, at this branch, through this channel. If your decisions only ever need what the till alone knows, the POS reports are enough; multi-branch, multi-channel operators outgrow that quickly.
Can we compare aggregator and direct channel economics without publishing our commission terms?+
Yes — the comparison is internal and stays internal. Your own settlement files and contracts carry the costs each channel imposes; the analytics apply them to each order privately, so the dashboard shows what each channel genuinely nets per branch without any term leaving your systems. Nothing is benchmarked against other operators and nothing is published. The point is not to know the market's terms; it is to know your own mix — which orders each channel carries, what each genuinely costs you to serve, and whether the mix is moving in the direction you chose. That reading tends to change behaviour at renewal time more than any negotiation tactic does.
What data has to exist before menu engineering is possible?+
Three things, and most operators have two of them already. Item-level sales with timestamps come from the POS. A cost per dish comes from recipes — even approximate ones, kept honest and updated when supplier prices move. And a mapping between the names each channel uses for the same dish, because a menu that appears under different item names on the website, the app and the aggregators cannot be summed. The recipe work is usually the missing piece, and it does not need to be perfect to be useful: a menu costed roughly and read by contribution beats a menu costed nowhere and read by volume, and the precision improves with each month it is used.
In this cluster
Next step
See your numbers in one picture
A data review that inventories what your systems already record and shows the first honest picture they could produce together.
See your numbers in one picture
What happens next
- 1We reply within one business day.
- 2A 30-minute scoping call with a senior consultant.
- 3A written proposal with scope, team, and timeline.
Or reach us directly

