BEE LOGICALExecutive Intelligence
Live · synced 42s ago
Net Revenue · MTD
₹1.25 Cr
▲ 12.4%
Shopify ₹74.9L · Amazon ₹49.9L
Net Contribution
₹27.7 L
▲ 8.1%
Margin 22.2% · after ad spend
Blended CAC
₹480
▲ 6.2%
5,200 new customers · MER 5.0×
Blended ROAS
4.2×
▲ 0.3×
Ad spend ₹24.96L
COD RTO Rate
18.5%
▼ 2.1pt
COD share 55% · target <12%
Stock Blocked in Transit
₹18.4 L
▲ 9.0%
142 shipments · incl. RTO return leg
Settlements Pending
₹22.1 L
● stable
Razorpay T+2 · Amazon disbursement
Orders · MTD
8,320
▲ 10.9%
AOV ₹1,500

Revenue & Contribution Trend 30 days

Daily net revenue vs. net contribution — the two lines a CEO watches diverge or track together.

Payment Mix

COD vs prepaid — the driver behind RTO and blocked capital.

Profit Waterfall ₹ Lakhs · MTD

Where every rupee of revenue goes — net revenue stepped down to true contribution.

Revenue by Channel

Storefront vs marketplace — one canonical number, two sources.

Marketing Performance by campaign

Spend, return and acquisition cost per campaign across Meta & Amazon Ads.
CampaignSpendRevenueROASCAC

ROAS by Channel

Efficiency at a glance — where the next rupee should go.

RTO by Courier

Return rate varies sharply by partner — route around the worst.

Highest-Risk Pincodes & Loss-Making SKUs

The specific rows draining margin right now.
ItemTypeSignalMargin impact

AI Action Center

6 actions ranked by ₹ impact
Not a chart — a decision queue. Each card is a threshold crossed, the recommended move, and the projected rupee impact.

Metrics & Formulas

Every number on the dashboard is defined once, here, and computed on the canonical layer — so it is identical for this client and for any client on a different stack. This is the reference that ends "the formula changed again."

Revenue

MetricFormulaSourcesRefreshViz
Gross RevenueΣ (unit price × qty) before discountsOrder lines (storefront + marketplace)~real-timeKPI
Net RevenueGross − Discounts − Refunds/Returns − CancellationsOrders, refunds~real-timeKPI + line
Revenue by ChannelNet Revenue partitioned by channel (Shopify / Amazon)Canonical Order.channel~real-timeDonut
AOVNet Revenue ÷ OrdersOrders~real-timeKPI

Profitability

MetricFormulaSourcesRefreshViz
COGSΣ (unit cost × qty sold)Product cost masterdailyWaterfall
Gross MarginNet Revenue − COGS · % = GM ÷ Net RevOrders, cost masterdailyKPI
Contribution MarginNet Rev − COGS − Shipping − Payment/COD fees − Ad spend − (RTO prob × RTO cost) − Returns handling − PackagingAll sourceshourlyKPI + waterfall
Contribution %Contribution Margin ÷ Net RevenueDerivedhourlyKPI
Profit per OrderContribution Margin ÷ OrdersDerivedhourlyKPI

Marketing

MetricFormulaSourcesRefreshViz
Ad Spend (blended)Σ spend across Meta + Amazon Ads (+ Google)Ad platform APIshourlyKPI
Blended CACTotal Ad Spend ÷ New CustomersAd spend, first-order flaghourlyKPI + line
ROASAttributed Revenue ÷ Ad Spend (blended & per channel)Ads, orders (attribution)hourlyBar
MERTotal Net Revenue ÷ Total Ad SpendOrders, adshourlyKPI
New vs ReturningRevenue split by first-order flagCustomer, ordersdailyBar / donut

RTO & Fulfilment

MetricFormulaSourcesRefreshViz
COD RTO %COD RTO orders ÷ COD shipped ordersShipment status~real-timeKPI + gauge
NDR %NDR orders ÷ shipped ordersShipment status~real-timeKPI
RTO CostΣ (fwd freight + reverse freight + handling + blocked-inventory carrying) · ~₹150–300/RTOShipping, cost masterdailyWaterfall
RTO by Courier / PincodeRTO % partitioned by courier / pincodeShipment~real-timeBar / table
On-time Delivery %Delivered ≤ promise ÷ DeliveredShipment SLAdailyKPI

Inventory & Cash

MetricFormulaSourcesRefreshViz
Stock Blocked in Transit (₹)Σ (units in forward-transit + RTO return-leg × unit cost)Shipment, cost master~real-timeKPI + table
Days of Inventory CoverSellable units ÷ avg daily units soldInventory, ordersdailyTable
Settlements Pending (₹)Captured-not-settled + marketplace payouts pendingGateway settlement, marketplace disbursementdailyKPI
Cash Conversion (days)Avg (settlement date − order date)Orders, settlementdailyKPI

Architecture & the Stack-Agnostic Layer

The reason a client on WooCommerce + Delhivery + Cashfree gets the same numbers as one on Shopify + Shiprocket + Razorpay: every source is normalized into one canonical model, and every formula runs on that model — never on the raw source. Swap a tool, keep the numbers.

1

Connector Layer

source adapters — normalize raw data → canonical events
StorefrontShopify · Woo · BigCommerce
ShippingShiprocket · Delhivery · Bluedart
PaymentsRazorpay · Cashfree · PayU
MarketplaceAmazon · Flipkart
AdsMeta · Google · Amazon Ads
2

Canonical Metric Layer

single source of truth — formulas computed once
Order / OrderLinerevenue, discounts, channel
Shipmentstatus, RTO, NDR, freight
Settlementfees, payouts, timing
AdSpendchannel, campaign, clicks
Inventoryunits, cost, in-transit
Customerfirst-order flag, LTV
3

Presentation & AI Layer

executive dashboard + AI action engine
Executive DashboardKPIs, trends, waterfall, mix, tables
AI Action Enginethresholds → ranked action cards
Why this fixes the Excel problem: today a metric is a formula living in someone's spreadsheet, re-typed and re-broken each month. Here the formula lives in the canonical layer, versioned and tested once. Real-time freshness comes from webhooks where a source offers them (orders, shipment status) and scheduled pulls where it doesn't (ad spend hourly, settlements daily).

Stack-agnostic mapping — same canonical entity, any tool

Canonical entityOption AOption BOption C
OrderShopify Admin APIWooCommerce RESTBigCommerce Orders
ShipmentShiprocketDelhiveryBluedart / Ecom Express
SettlementRazorpayCashfreePayU
Marketplace OrderAmazon SP-APIFlipkart Seller API
AdSpendMeta Marketing APIGoogle Ads APIAmazon Ads API
Onboarding a new client is therefore a connector-mapping exercise, not a rebuild: point their tools at the matching adapter, and the entire dashboard, every formula and every AI action works unchanged.

AI Action Engine

The dashboard shows what happened; the engine says what to do about it. Version one is a deterministic rules engine over the canonical metrics — using Bee Logical's own KPI thresholds — that ranks recommendations by projected rupee impact and surfaces them as action cards. It graduates to ML as data compounds; it never needs a cold start because the rules are the framework.

Rule library — signal → action

SignalConditionRecommended actionProjected impactData
Budget reallocationROAS(A) > blended×1.5 AND ROAS(B) < 1.0Shift spend from B → A, up to A's saturation point+ incremental revenueAdSpend, Orders
Scale winnerROAS > target AND spend < cap AND CM > 0Increase budget +15–20% on the campaign+ profitable revenueAdSpend, CM
Creative fatigueCTR ↓ >30% over 7d AND frequency > 3Refresh creative / rotate ad set− rising CPPAd platform
RTO pincodeCOD RTO(pincode) > 30%Switch pincode cluster to prepaid-only+ RTO cost savedShipment
SKU margin leakContribution %(SKU) < 0Review price or COGS on the SKUstop the bleedOrders, cost
Inventory reorderDays cover < supplier lead timeReorder now to avoid stock-outavoid lost salesInventory
Blocked capitalNDR age > 5d OR in-transit value risingTrigger reattempt / expedite backlogunlock ₹ tied upShipment
Margin alertContribution % < target (e.g. 15%)Surface the top driver of the dropprotect marginDerived
How a card is built: a rule fires on the canonical metrics → the engine quantifies the rupee impact from the same data → cards are ranked highest-impact first → the CEO acts (or dismisses). Because every rule reads the canonical layer, the engine is stack-agnostic and identical across clients. The roadmap layers ML (anomaly detection, spend-response curves, RTO probability) on top of this rules base as volume grows.

Bee Logical · Executive Intelligence Dashboard — prototype · figures are illustrative sample data for a ~₹1.25 Cr/month brand on Shopify + Shiprocket + Razorpay + Amazon + Meta/Amazon Ads · charts require an internet connection to render.