An owner shouldn't have to assemble reports
to understand what is happening in the chain
Owner Intelligence is a management loop for a chain: POS, loyalty, delivery, costs and reputation come together in one pane. A digest arrives every morning, per-location anomalies find you on their own, and an advisor works on top of the data — explaining the numbers in plain words and proposing decisions on menu margin, load, control and marketing.
— 01 / What breaks in a chain
Management attention is the owner's scarcest resource — and it is spent assembling spreadsheets
The problem isn't missing data. The data exists — in the POS, in loyalty, at the aggregators, in the bank, in managers' chats. Nobody consolidates it by morning and nobody explains it.
To understand how things are going you ask three people and open five spreadsheets
→ One pane: revenue, checks, average check, locations, channels, categories — for any period, against the previous one
The report arrives a week late and no longer decides anything
→ Data loads itself; the digest is in your messenger before opening hours — for yesterday
The problem is visible once it is already in the P&L
→ A location's deviation is measured against its own median for that weekday and arrives as an alert the same day
Discounts, voids, deletions and bonus giveaways live outside your field of view
→ Control and fraud radar: the norm is computed on your own data, deviations are shown by operation, guest and shift
The menu is discussed by taste, not by margin
→ Every item plotted on demand × margin, with a concrete action: hold, raise price, promote, revise
Aggregators take their share and nobody calculates how much exactly
→ Commissions at your contracted rates, city by city, and the amount leaving through each channel
An analyst costs a full-time salary and still explains the numbers once a month
→ The advisor works on your numbers continuously, answers a question in text, and translates jargon into plain language
— 02 / Who buys this
One loop — four different answers to «why»
Management data is needed by four functions at once, each in a different cut. Here is what changes for each, and the metric that shows it.
Owner · founder
TodayTo understand how things are going, they ask three people and still aren't sure.
With the loopA morning read and signals, unprompted. A question in plain words instead of a task for an analyst.
MetricHours of attention per week and time from event to decision
CEO · COO
TodayAccountable for a result they see a week late, in a different format for every country.
With the loopOne picture across all locations, norms computed on your own data, and anomalies that arrive on their own.
MetricReaction speed to a deviation and share of locations performing at their own norm
CMO · marketing director
TodayChannels are judged by spend and turnover. What a channel yields after commissions stays unanswered.
With the loopRevenue and margin by channel, real aggregator commissions at your contracted rates, promo effect against the weekday norm, loyalty and churn.
MetricChannel contribution to margin rather than turnover; retention and visit frequency
Commercial director · service director
TodayMenu, prices and shifts are argued from taste and experience, because the data isn't at hand.
With the loopDemand × margin per item with a ready action, hourly load, cost, and operations control.
MetricBasket margin, food cost against the median, and losses to discounts and write-offs
— 03 / Who it is for
Chains where decisions cost money and data is scattered across systems
The loop is built where there are several locations, guest flow, and several data sources: POS, loyalty, delivery, costs. The more locations and jurisdictions, the larger the effect — because that is exactly the configuration in which an owner stops seeing the whole business.
Markets and currencies
Moldova, Romania, Ukraine are the priority markets. Multi-currency by default: Moldovan leu, Romanian leu, hryvnia, forint and euro are consolidated into one comparable picture when the chain operates across countries and each jurisdiction keeps its own POS separately.
— 04 / Chain of events
An owner's day with the loop
Not «a dashboard you have to log into», but a sequence that reaches you on its own. You log in only when you want to dig.
Data is pulled
The loop collects checks, payments, channels, loyalty and platform ratings, computes weekday medians and looks for deviations. Nothing to click.
Digest
Yesterday: revenue, checks, average check, comparison with the norm for that weekday, best and weakest location, top items. One message instead of a round of managers.
Signals
«This location is 37% below its own norm», «average check up 12% at another» — each line links to the location breakdown: dynamics, hours, items, history.
A question in words
«Why did revenue drop in this city last week?» — an answer with numbers, broken down by channel and location. No SQL, no exports, no analyst.
Structural review
What changed structurally, not for a day: channels, menu, hourly load, loyalty, reputation. Recommendations in money: «food cost 33% against a 26% median — roughly N per month overpaid».
P&L and decisions
P&L by location, margin, a profitability ranking of locations, where the loss is structural and where it is one-off. Exactly the cut usually prepared for a board meeting.
Control
Fraud radar and giveaways: suspicious bonus accruals and redemptions, abnormal discounts and deletions, operations outside working hours — linked to guest and shift.
— 05 / What's inside
Twelve modules in one pane
The composition is configured per chain: we connect what has data; the rest is switched on later without rebuilding the loop.
Overview
Key metrics with trend, revenue by day, best and weakest day, share by location and city.
Countries and cities
Markets compared in one currency, with the channel structure inside each — you immediately see who lives on dine-in and who on aggregators.
Channels
Dine-in, pickup, own delivery, aggregators, in-app payments. Separately — how much leaves as commission at your rates.
Menu and margin
Demand × margin per item, an action for each group, a separate card for add-ons and sauces, category shares inside kitchen and bar.
Forecasts
Expected revenue and load based on the location's history and the weekday — so shifts and purchasing aren't set on gut feel.
Anti-fraud
Accrual and redemption norms computed on your own data, and deviations from them: not «suspicious in general» but «above your own norm».
Fraud radar
Operations outside working hours, checks paid entirely with bonuses, repeating patterns around one guest or one shift.
Operations control
Deletions, voids, discounts, operations with no stated reason — with amounts and locations. Plain language instead of POS jargon.
Reputation
Ratings by platform and location, dynamics, loyalty index. Docks with Reputation Loop into a single contour.
Load
Hour × day heat map, actual opening and closing times from the first and last check, peak hours per location.
Loyalty and guests
RFM segments, cohorts, churn, guest card, bonus transactions for the period. Staff cards with no real visits are visible separately.
Cost and stock
Food cost and margin per item, deviations by location, candidates for recipe and purchasing review.
— 06 / Alternatives
Three ways to get the management picture
Excel is honest but late. A BI contractor builds what the brief says, then leaves. The loop lives with the chain and explains itself.
| Reports and Excel | BI contractor | Owner Intelligence | |
|---|---|---|---|
| Who consolidates data | managers and the accountant, by hand | a developer, per written brief | the loop, daily and unprompted |
| Speed | a week, sometimes a month | days, after each change request | next morning |
| Who explains the numbers | nobody | an analyst, billed separately | the advisor in the dashboard and in your messenger, in your language |
| Sources | POS and spreadsheets | whatever got connected during the project | POS, loyalty, delivery, costs and reputation in one loop |
| Anomalies | noticed through revenue | only if specified in the brief | computed against the location's own norm and delivered to you |
| Cost of growth | new location — new spreadsheet | new location — new sprint | new location — a line in the configuration |
| Multiple countries | consolidated into euro by hand | rare and expensive | multi-currency and multi-jurisdiction by default |
| Who owns it | it lives in people's heads | with the contractor, along with the code | your server, your data, documentation handed over |
— 07 / Economics
Counted in the open: where the money actually sits
An illustrative calculation with visible assumptions. Every input is listed separately: substitute yours and see whether it holds.
Assumptions
- Chain size
- 10 locations
- Chain turnover
- €400,000 per month
- Owner's attention
- 1 hour a day assembling numbers
- Typical deviation
- one location 15% below its norm
One hour a day of the person at the top
a quarter of working time spent assembling numbers rather than deciding on them
Food cost 3 points above the median at three locations
visible only once cost sits next to demand, item by item
A location 15% below norm, caught a month later instead of the next morning
revenue lost on a single event; there are several such events in a year
Aggregator commissions at contracted rather than averaged rates
the difference that decides where a channel is worth keeping and where it eats the margin
A €6,500 Full implementation pays for itself on a single deviation caught in time. The first run in a real sixteen-location chain surfaced three — not a promise of results, just what the data shows once someone connects it.
Illustrative figures at roughly €40,000 monthly turnover per location. In the review we run them on your data and say plainly which sources are missing for the layers you want.
— 08 / Proof
A loop that reaches the owner every morning
Under NDA · food service chain, four countries
16 locations, four jurisdictions, one management loop
Each country ran its POS as a separate, unlinked network; loyalty lived in a third system and the aggregators in a fourth. We built the loop: overnight data collection, a 9:00 digest, anomalies against the weekday median, P&L by location, the menu plotted on demand × margin, and a control module. The very first run surfaced three real per-location deviations; the menu breakdown showed that nearly half of revenue comes from items priced below the margin norm. Technical specifics are under NDA.
Open the case— 09 / Pricing
From €3,500 for a loop that replaces the reporting routine
Implementation is fixed and depends on the number of sources and the depth of layers. After that: operation, or full support with a human alongside.
Core
implementation, one-off
Up to 10 locations, one market. POS connection, single pane, morning digest, per-location anomalies, questions in words, location breakdown. Further layers by priority.
Discuss implementationFull
implementation, one-off
Multiple countries and currencies. Everything in Core plus P&L by location, menu and margin, channels and aggregator commissions, loyalty with RFM and cohorts, operations control and fraud radar, reputation.
Discuss implementationAdvisor + External CMO
per month
The loop plus a human: a weekly review of the numbers, priorities for the month, decisions on margin, channels and marketing. This is the standard External CMO format — run on your data instead of slides.
Discuss implementationOperation without support — €350 per month: server, models, source updates, adding locations. Not included: keeping your primary documents and bookkeeping. The loop computes what you actually record; what's missing, we'll show you in week one.
— 10 / Objections
What usually gets asked before signing
«We already have POS reports and BI»
The POS shows its own part: it doesn't know aggregator commissions, doesn't consolidate several countries into one currency, doesn't compute a weekday norm for a specific location, doesn't see loyalty or reputation, and explains nothing. The loop doesn't replace the POS or argue with BI — it connects what they don't, and adds the explanation layer.
«Our IT team could build this»
They could — collecting data isn't magic. The difference is what comes after: someone has to calibrate norms, fix sources when they change, hold the quality of explanations, and add layers. We hand over a finished loop with documentation: if IT wants to take over operation, that's available at any point.
«We don't want vendor dependency»
The loop lives on your server, the access and the data are yours, documentation is handed over. This is an implementation you own, not a subscription that switches off together with access to your own numbers.
«Guest data and commercial confidentiality»
What goes to the model are aggregates for explanations and recommendations, not personal data. Phone numbers and identifiers are hashed at collection. Access is separated: a manager sees their location, the owner sees the chain.
«The team won't use it»
Which is why the primary format is not a dashboard but a digest and alerts in the messenger the team already lives in. The dashboard is opened when someone wants to dig. If people have to be forced to use a tool, the tool is designed wrong.
«What if we don't have enough data?»
That surfaces in week one, not two months in. You get a source map and an honest list of what's missing for the layers you want: a full P&L needs costs and staffing, and if those don't exist we say so before you pay for a layer that cannot be built.
— 11 / Implementation
Four steps to the first digest
Week 1 · Sources
We look at what the POS, loyalty, aggregators and bookkeeping actually expose. Output: a source map and an honest list of what is missing for the layers you want.
Weeks 2–3 · The loop
Data collection, single pane, digest and anomalies — on your data, not on a demo. The first digest arrives before implementation ends.
Week 4 · Calibrating norms
We tune what counts as a deviation for you specifically: thresholds, minimum amounts, promo periods. Without this step a loop screams at campaigns and stays silent on real problems.
Then · Layers by priority
P&L, menu and margin, control, loyalty, reputation — switched on one at a time, in the order of your pain, not our architecture.
— 12 / Questions
What chain owners ask
Which POS and systems can be connected?+
One practical criterion: if the system exposes an API or a regular export, it connects. Already live: Syrve and iiko, loyalty programs with their own API, delivery platform storefronts, payment and QR services, cost and staffing spreadsheets. For retail and e-commerce: inventory systems and store platforms.
Is this a BI dashboard or an AI advisor?+
Two layers of one product. The lower layer is data: collection, consolidation, norms, anomalies, cuts. The upper layer is the advisor: it answers a question in words, explains the deviation and proposes an action with an amount. Without the lower layer the upper one hallucinates — which is why we don't sell «a chat on top of Excel».
Does our data leave our systems?+
The loop is deployed on your server or on one dedicated to you. What goes to the model are aggregates for explanations and recommendations, not guests' personal data. Phone numbers and identifiers are hashed at collection.
Why do we need this if the POS already has reports?+
The POS shows its own part: it doesn't know aggregator commissions, doesn't consolidate several countries into one currency, doesn't compute a weekday norm for a specific location, doesn't see loyalty or reputation, and explains nothing. The loop doesn't replace the POS — it assembles what the POS never connects.
How many locations make this worthwhile?+
Three to four and up. With one location the owner holds the picture in their head. From the third, a gap opens between what happens and what they learn — and that gap is closed by a loop, not by hiring another manager.
Who uses it besides the owner?+
The operations director, the finance lead, location managers — with different access rights. The owner sees the whole chain, a manager sees their location. Digests and alerts are configured per recipient.
What if our bookkeeping is partial and some data is on paper?+
We start with the layer that has data: revenue, checks, channels, locations. A full P&L needs costs and staffing — if those don't exist, we say so in week one, not two months into implementation.
How is this different from hiring an analyst?+
An analyst costs more as an employee, works on request, goes on holiday and takes the context along. The loop runs every morning and doesn't depend on one person. Where human interpretation and decisions are needed, we add them in the External CMO format — but on top of numbers that are already there.
Next step
We'll work out on your data which three signals the loop shows in week one
One meeting: what sources the chain has, what can be assembled from them in two weeks, and which decisions that changes in the first month. Output: a source map — whether or not we work together afterwards.