What the business needs you to decide this fortnight.

as of 2026-08-04

Part of this briefing is running on incomplete inputs — the affected figures are shown empty and labelled, never estimated

· calc_leakage is not present in analytics.duckdb

Briefing

Over the next 14 days, PKR 4,91,720 of gross margin is at risk across 28 territory and SKU lines in 26 territories — stock that will run out before replenishment can reach it, and orders that are already late. Demand over the same two weeks just past ran 11.4% lower than the fortnight before it. A further PKR 16.65m of commercial spend this year cannot be tied to a promotion, so nothing can be said about what it returned. Separately — and never added to those figures — PKR 133.31m of cash is sitting in stock that has more than 120 days of cover. That is a balance sheet problem, not a profit one, and it is fixed by clearing stock rather than by protecting margin. Against all of that, Genaima has put PKR 1.01m of protected margin in front of a decision-maker across 25 open cases, PKR 2,85,243 has been approved, and PKR 1,13,143 has been executed and measured. The distance between those three numbers is the honest picture of how much of this the organisation is actually converting.

Read the full briefing

The six numbers

every figure opens its own calculation

Gross margin at risk

PKR 4,91,720

P&L · flow

next 14 days

24 lines short of cover · 4 waiting on a late order

How this is calculated

Formula

margin at risk = Σ (demand over the next 14 days − stock on hand − stock already inbound) × margin per unit, over every territory × SKU line that cannot cover the horizon; plus, for lines with a purchase order past its promised date, 14 days of demand × margin per unit

Inputs and values

Horizon: 14 days from 2026-08-04
Lines short of cover: 24
Margin at risk from availability: PKR 2,68,841
Lines with a late purchase order: 4
Margin at risk from late supply: PKR 2,22,878
Margin per unit: observed selling price × 0.45

Result

PKR 4,91,720

Sources

What this does not tell you

Demand is the last 28 days of observed sell-through per line, not a forecast.
Margin per unit is derived from observed selling price at a 45% gross margin, because item cost is not present in any connected feed.

Excess and ageing stock

PKR 133.31m

Cash · balance

held at 2026-08-04

3,628,419 units across 583 lines with over 120 days of cover

How this is calculated

Formula

trapped cash = Σ units on hand × cost per unit, over every territory × SKU line holding more than 120 days of cover, where cost per unit = observed selling price × 0.55

Inputs and values

Cover threshold: 120 days
Lines above threshold: 583
Units held: 3,628,419
Value at cost: PKR 133.31m
Value at realised price: PKR 242.37m

Result

PKR 133.31m

Sources

What this does not tell you

FR-C2: this is one-time cash already spent and sitting in stock. It is never added to the P&L figures on this page.
Recovering it requires markdown, so the cash released will be lower than the value shown here.

Marketing spend with no traceable return

PKR 16.65m

P&L · flow

year to date · 211 days

26 campaigns whose return cannot be measured

How this is calculated

Formula

unprofitable expenditure = Σ campaign spend where the campaign cannot be tied to a trade programme in the distribution model, so no uplift can be attributed to it

Inputs and values

Total commercial spend: PKR 96.37m
Of which unattributable: PKR 16.65m
Campaigns affected: 26
Window: 211 days to 2026-08-04

Result

PKR 16.65m

Sources

What this does not tell you

calc_leakage is not present in this build, so the figure is computed from campaign attribution instead. It measures spend whose return is unmeasurable, not spend proven to be wasted.
Attribution requires a campaign identifier that resolves to a promotion. Spend that cannot be resolved is reported here rather than assumed profitable.

Value awaiting your decision

PKR 1.01m

Pipeline

trailing 90 days

25 cases waiting on a decision

How this is calculated

Formula

recommended value = Σ expected value of the recommended option on every case the system has put in front of a decision-maker, at the ledger stage "recommended", in protected margin only

Inputs and values

Ledger stage: recommended
Value type: protected margin
Amount: PKR 1.01m
Cases awaiting a decision: 25

Result

PKR 1.01m

Sources

What this does not tell you

Different value types are different currencies and are never added. This figure is protected margin only.
Recommended is not approved. The gap is shown in the value ledger strip below.

Approved, being carried out

PKR 2,85,243

Pipeline

trailing 90 days

PKR 2,33,703 has reached execution

How this is calculated

Formula

approved value = Σ expected value recorded at the ledger stage "approved", written only when a named approver accepts a case

Inputs and values

Ledger stage: approved
Amount approved: PKR 2,85,243
Amount recorded as executed: PKR 2,33,703
Cases decided: 14

Result

PKR 2,85,243

Sources

What this does not tell you

Approval is recorded by the decision endpoint, so this figure moves only when a human actually approves a case.
Execution is recorded separately; approved value is not yet realized value.

Realised and finance-verified

PKR 1,13,143

Pipeline

trailing 90 days

of which finance has signed off PKR 53,660

How this is calculated

Formula

realized value = Σ recorded at the ledger stage "realized"; finance-verified is the subset that carries a named verifier and a timestamp. The two are shown as a headline and a subset, never added together

Inputs and values

Realised: PKR 1,13,143
Finance-verified: PKR 53,660
Baseline: the case baseline recorded when the case was opened

Result

PKR 1,13,143

Sources

What this does not tell you

Realized value is measured against the baseline captured when the case opened.
A finance-verified amount cannot be written without a named verifier.

These numbers are not added together.why a flow never joins a balance

Margin at risk and unprofitable commercial spend are profit-and-loss leakage — money that will not be earned in a period. Excess and aging inventory is cash already spent and frozen in stock. Combining a flow with a balance produces a headline that is arithmetically valid and economically meaningless, so Genaima reports them side by side and never sums them.

What happened to the value we claimed

protected margin · trailing 90 days to 2026-08-04

1Identified
PKR 1.55m
100.0% of identified
2Recommended
PKR 1.01m
64.7% of identified
65% survived this step
3Approved
PKR 2,85,243
18.4% of identified
28% survived this step
4Executed
PKR 2,33,703
15.0% of identified
82% survived this step
5Realised
PKR 1,13,143
7.3% of identified
48% survived this step
6Finance-verified
PKR 53,660
3.5% of identified
47% survived this step

Read this left to right. The drop between two stages is value that did not survive the step — identified but not worth recommending, recommended but not approved, approved but never executed. A ledger that reported only the last number would be marketing.

How each stage is calculated

Formula

each stage = Σ protected margin recorded at that stage of the value ledger. The percentage is that stage as a share of what was identified, so the drop between two bars is value that did not survive the step

Inputs and values

Identified: PKR 1.55m · 100.0% of identified
Recommended: PKR 1.01m · 64.7% of identified
Approved: PKR 2,85,243 · 18.4% of identified
Executed: PKR 2,33,703 · 15.0% of identified
Realised: PKR 1,13,143 · 7.3% of identified
Finance-verified: PKR 53,660 · 3.5% of identified

Result

PKR 53,660 finance-verified

Sources

What this does not tell you

Protected margin only. The four value types are different currencies and the ledger refuses to total across them.
Every entry in this ledger is written by the case build or the decision endpoint with the demo clock date, so the 90-day window contains the whole ledger; no entries are excluded by the window.

The next 14 days, if nothing is done

forward exposure by stream · 2026-08-04 → 2026-08-18

Forward exposure by stream over the next 14 days
Stream	Exposure if nothing is done	Awaiting a decision	Expected value if taken
Stock availability Lines that will run out before replenishment can reach them How stock availability is calculated Formula Σ (demand over the next 14 days − on hand − inbound) × margin per unit, over lines that cannot cover the horizon Inputs and values Lines 24 Exposure PKR 2,68,841 Open cases 21 Expected value of recommended actions PKR 5,92,821 Result PKR 2,68,841 Sources Derived stock signals`analytics.duckdb · signal_stock` Decisions`state.db · decision_case` What this does not tell you Exposure is what happens if nothing is done. Expected value is what the recommended action is priced to recover, so the two are not equal by design. Only cases the governance layer actually recommended are counted. An option that policy blocked is priced but never offered, so it is excluded from both columns rather than quoted as available value.	PKR 2,68,841	21 open cases awaiting a decision	PKR 5,92,821
Supply reliability Purchase orders already past the date the supplier promised How supply reliability is calculated Formula Σ 14 days of demand × margin per unit, over lines with an overdue purchase order Inputs and values Lines 4 Exposure PKR 2,22,878 Open cases 0 Result PKR 2,22,878 Sources Purchase order feed`source_registry · erp_purchase_orders.csv` Derived stock signals`analytics.duckdb · signal_stock` What this does not tell you An overdue order is assumed not to arrive inside the horizon, because no connected feed carries a revised delivery date.	PKR 2,22,878	0 detected, no case opened yet	PKR 0
Commercial spend Spend that cannot be tied to a promotion, pro-rated to the horizon How commercial spend is calculated Formula unattributable spend year to date ÷ days elapsed × 14, giving the share of unmeasurable commercial spend that falls inside the horizon at the current rate Inputs and values Unattributable spend, year to date PKR 16.65m Days elapsed 211 Horizon 14 days Pro-rated exposure PKR 1.10m Result PKR 1.10m Sources Marketing platforms feed`source_registry · campaign_spend.csv` Unmapped records`bronze_campaign` What this does not tell you This is a pro-rated run rate, not a forecast. It assumes commercial spend continues at the rate observed so far this year.	PKR 1.10m	0 no marketing actions registered in this build	PKR 0
Total profit-and-loss exposure P&L streams only — trapped cash is not added in (FR-C2) How the total is calculated Formula total = sum of the P&L streams above only. Aging inventory is trapped cash and is excluded from this total by rule (FR-C2) Inputs and values Stock availability PKR 2,68,841 Supply reliability PKR 2,22,878 Commercial spend PKR 1.10m Aging inventory (excluded) PKR 133.31m Result PKR 1.60m Sources Derived stock signals`analytics.duckdb · signal_stock` Marketing platforms feed`source_registry · campaign_spend.csv` Decisions`state.db · decision_case` What this does not tell you Annual P&L leakage and one-time trapped cash are different kinds of money and are never added together.	PKR 1.60m	21	PKR 5,92,821
Aging inventoryreported separately Cash already spent and sitting in stock over 120 days old How aging inventory is calculated Formula Σ units on hand × cost per unit, over lines with more than 120 days of cover Inputs and values Lines 583 Units 3,628,419 Value at cost PKR 133.31m Result PKR 133.31m Sources Derived stock signals`analytics.duckdb · signal_stock` What this does not tell you FR-C2: a balance, not a flow. Excluded from the board total by rule, not by oversight.	PKR 133.31m	0 markdown actions not yet opened as cases	not comparable

The value column counts only what is still awaiting a decision. A further PKR 4,13,799 was recommended in the same window and has since been decided, which is why the ledger strip below opens on a larger number.

Board-level sizing

arithmetic on the revenue base — not a forecast of capture

Sizing exercise

Sizing only — not a forecast of capture

How large would a 1–3% improvement on revenue be? This is arithmetic on the revenue base, nothing more. Nothing on this side of the page has been modelled, recommended, or approved.

1%2%3%

Sizing only — not a forecast of capture. Genaima has not claimed it can capture any of this.

How the sizing is calculated

Formula

scenario = net sales over the stated window × 1%, 2% or 3%

Inputs and values

Revenue base: PKR 2.32bn
Window: net sales observed 2026-05-06 → 2026-08-03
1% of revenue: PKR 23.24m
2% of revenue: PKR 46.47m
3% of revenue: PKR 69.71m
Modelled and recommended: PKR 1.01m

Result

1% = PKR 23.24m · 2% = PKR 46.47m · 3% = PKR 69.71m

Sources

What this does not tell you

This is arithmetic on revenue, not a model output. It answers "how big would a 1–3% improvement be", nothing more.
It is deliberately not comparable with the modelled figure beside it: one is a scale, the other is a set of priced, approvable actions.

Modelled capture

PKR 1.01m

What Genaima has actually modelled and put in front of a decision-maker: priced options on real cases, each with a named approver, an expected value and a baseline it will later be measured against.

This is the only figure on this page Genaima stands behind, and it is the one tracked through the ledger above. It is deliberately not comparable with the sizing exercise beside it.

Open the decision queue

Recently executed actions

8 actions · every decision carries the name of the person who made it

closed

Restock approved — Cherry 1.5L PET 146 at ABC Distribution Faisalabad Territory 1

system:orchestrator · 2026-07-17 · case closed after outcome verification

PKR 20,651

expected value

View case →

closed

Restock approved — Berry 4-Pack 436 at ABC Distribution Faisalabad Territory 1

system:orchestrator · 2026-07-09 · case closed after outcome verification

PKR 22,349

expected value

View case →

closed

Restock rejected — Apple 6-Pack 362 at ABC Distribution Lahore Territory 6

system:orchestrator · 2026-07-03 · case closed after rejection

PKR 22,657

expected value

View case →

closed

Restock approved — Original 4-Pack 437 at ABC Distribution Lahore Territory 7

system:orchestrator · 2026-07-01 · case closed after outcome verification

PKR 27,527

expected value

View case →

closed

Restock rejected — Guava 200ml Pack 361 at ABC Distribution Islamabad Territory 4

system:orchestrator · 2026-06-25 · case closed after rejection

PKR 22,408

expected value

View case →

closed

Restock approved — Citrus 12-Pack 428 at ABC Distribution Islamabad Distribution Hub

system:orchestrator · 2026-06-23 · case closed after outcome verification

PKR 32,433

expected value

View case →

closed

Restock approved — Summer 6-Pack Bundle 498 at ABC Distribution Lahore Territory 2

system:orchestrator · 2026-06-19 · case closed after outcome verification

PKR 38,934

expected value

View case →

closed

Restock rejected — Citrus 12-Pack 428 at ABC Distribution Lahore Territory 6

system:orchestrator · 2026-06-17 · case closed after rejection

PKR 27,088

expected value

View case →

Measured, not claimed

computed live from the connected data

2,461

territory × SKU pairs scored nightly

How this is calculated

Territory × SKU pairs scored nightly

Formula

count(*) over signal_stock — one row per territory × SKU pair the nightly analytics build scored; no sampling, no extrapolation

Inputs

signal_stock row count: 2,461

Value on this screen

2,461

Sources

signal_stock

≈15 days

warning before shelves empty · median cover of at-risk pairs

How this is calculated

Warning before shelves empty

Formula

median(days_cover) over signal_stock where days_cover < 28 — the median days of cover remaining across pairs already inside the stockout window, i.e. how much notice the queue gives before stock reaches zero

Inputs

days_cover of pairs with cover below the 28-day floor — read per row from the source tables, not re-stated here

Value on this screen

≈15 days

Sources

signal_stock

headline figures · all 10 rendered on screen open their calculation · 2 published only in the API and one-pager

How this is calculated

Headline figures with a registered calculation

Formula

FIGURE_IDS.length — the number of entries in the FR-F4 calculation-panel registry; runParity() recomputes every one from its source tables by an independent path and drift fails the build

Inputs

FIGURE_IDS: 12 registered ids

Value on this screen

Sources

FIGURE_REGISTRY — src/server/parity.ts

timed live on stage — not computed< 90 secondsfrom queue to audit trail · watch the clock, not this screen

What we do not claim

the refusals, stated deliberately

No forecast-accuracy promises on your business — accuracy on real data is unknowable until we run on your export, and synthetic results stay labelled synthetic.
Margin protection and cash release, not revenue growth. We protect what you already earn.
Not a headcount plan — the labour figure measures misallocation against traffic, nothing else.
Nothing executes without human approval. Every action above stops at a person.

Contract v1 · figures read from the analytics build and the decision state state at request time · synthetic distribution demonstration data.