Sign in
Set up for FMCG and CPG brands

AI for insights and analytics

AI analytics for FMCG and CPG brands

A sales director wants three things explained: where revenue came from, where share moved, and which promotions paid back. AI analytics reads shipments, market data, distribution and promotion files, splits every change into price, volume, mix, new and lost, and says which retailer and SKU did it.

Ask it in your own words

These are the questions CPG brands ask. Each one maps to a measure our tools compute from your files. The AI model chooses the measure and explains the result; the arithmetic is done by our code, and every total is checked.

The questionThe measure behind itWhat comes back
Explain where my revenue is coming from this quarter.Price, volume and mix bridgeThe change split into price, volume, mix, new and lost lines, by retailer and category, summing to the change exactly.
Where are we losing market share?Rate of saleShare by category and retailer from the market file against the prior period, with the SKUs behind each move.
Which promotions actually paid back?Promotional net liftLift against baseline, the dip after, net units and margin against trade funding, per event.
Where are we listed but not stocked, or stocked nowhere?Distribution voidsSKU and store combinations that similar stores carry and this one does not, valued at rate of sale.
Are we shipping more than the retailer is selling?Sell-in against sell-outShipments against retailer sales by SKU, with weeks of cover and the inferred out-of-stocks.
What are chargebacks costing us, and why?Chargebacks by retailer and reasonDeductions by retailer and reason code against invoiced value, with the repeat reasons.

A conversation, with the figures

Every figure below was computed by a tool from the rows and checked before it was shown. The lines sum; the percentages match; each line opens to the rows that make it.

What is driving the revenue change this quarter?

Net sales are up $5.7m against the prior year, split across 3 effects that sum to the change. Price is the largest at $3.1m, 54% of the movement.

Price$3.1m54% of the total
Volume$1.9m33% of the total
Mix$700k12% of the total
The 3 lines sum to$5.7m0 unexplained

Which retailers are behind the price effect?

3 retailers account for $2.9m of the $3.1m Price effect, 94% of it. The step in list price landed in March and held in all three.

Northway Grocers$1.4mprice up 4.1%, volume down 1.2%
Cityline Convenience$900kprice up 3.8%, volume flat
Meadow Markets$600kprice up 4.4%, volume down 2.9%
These 3 are$2.9m94% of Price

What should the category team look at next?

Volume at Meadow Markets, where the price step cost the most units, and the two promotions there that ran below baseline after the event. The bridge points at the retailer; the promotion table says which event.

Category managerReview Meadow Markets' post-promotion dip by SKUThis week
National account managerTake the price and volume split to the Northway reviewNext review
Sales directorPut the bridge in the monthly pack, by retailerMonthly
Explain where my revenue is coming from this quarter.Where are we losing market share?Which promotions actually paid back?Where are we listed but not stocked, or stocked nowhere?

The measures behind the answers

Each measure has one formula, one source and one meaning. They are computed per retailer and account manager and in total, and every one carries an identity that must hold before it is shown.

MeasureFormulaFromWhat it tells you
Range penetration by channelStores ranging the SKU ÷ stores in the channel that should, by the channel's own core rangeStore list; sell-out dataWhether the agreed range is on the shelf
Distribution voidsStores where similar stores sell the SKU and this one does not, valued at the median rate of saleSell-out data; store attributesThe list of store and SKU gaps worth closing
Sell-in against sell-outUnits shipped − units sold through, cumulative, per retailerShipment ledger; sell-out dataInventory building or draining in the channel
Weeks of cover at the retailerEstimated retailer inventory ÷ average weekly sell-outShipments; sell-out; opening stockWhether the next order will be large or small
Inferred out-of-stocksStore-days with zero sales where the SKU normally sells and similar stores soldDaily sell-out by storeLost sales at the shelf, by store and SKU
Promotional net liftPromo units − baseline − post-promo dip, per account, against the account's own baselineSell-out; promotions calendarWhat a promotion really added
Return on trade spendNet lift margin ÷ trade spend on the eventNet lift; deductions and promo fundingWhich events pay for themselves
Chargebacks by retailer and reasonDeductions by reason code ÷ sales, per retailer; disputed and recoveredDeductions fileProcess failures billed as penalties
OTIF under the retailer's ruleOrders on time and in full by the retailer's definition ÷ ordersOrder and delivery files; retailer termsThe score the retailer sees, not the one the brand reports
Rate of saleUnits per store per week, for stores ranging the SKUSell-out; store listPerformance independent of distribution

Each measure is worked through, with the export it comes from and what to drop, in Sales KPIs for CPG and consumer brands.

What the AI model does, and what our tools do

The AI model chooses the measure

From your question and the measures declared for FMCG and CPG brands, the model picks the one that answers it, and the period and comparison the question implies.

Our tools do the arithmetic

Deterministic code reads the rows, computes the measure, and checks the identities below. The same question on the same data gives the same answer, every time.

The AI model explains, and cites

The AI model writes the sentence around the result, naming the retailer or account behind it. It states no figure that is not in the result, and every figure links to its rows.

The identities that must hold

Sell-in and sell-outOpening retailer stock + sell-in − sell-out = closing stock
Net liftPromo units = baseline + gross lift; net lift = gross lift − post-promo dip
ChargebacksDeductions = valid + disputed open + recovered + written off
OTIFOrders = OTIF + late only + short only + both

What it reads

The exports CPG brands already produce. Column names are mapped once and the mapping is reused. A file is the way in; scheduled delivery and connections to your systems come with the plan, and every source is listed here.

  • Store list
  • Sell-out data
  • Store attributes
  • Shipment ledger
  • Shipments
  • Opening stock
  • Daily sell-out by store
  • Promotions calendar
  • Net lift
  • Deductions and promo funding
  • Deductions file
  • Order and delivery files
  • Retailer terms

Who owns each answer

An answer is a list with an owner and a cadence, or it is a chart nobody works.

MeasuresOwnerCadence
Inferred out-of-stocks; distribution voidsAccount managers; field salesWeekly
Sell-in against sell-out; weeks of cover; OTIFAccount managers with supply planningWeekly to monthly
Net lift; return on trade spendCommercial director; revenue managementAfter each event; quarterly
Chargebacks by reasonCustomer service and logistics, with financeMonthly

Questions CPG brands ask about AI analytics

Does the bridge really sum to the change?

Yes, and it is checked. Price, volume, mix, new and lost are computed from the shipment file by SKU and retailer, and the tool asserts that end minus start minus the sum of the lines is zero before it shows anything. If a line cannot be attributed it appears as 'other' with its value, never as a silent gap.

Where does market share come from?

From the market data you upload, whether that is a syndicated panel export or retailer EPOS. Share is your sales over the category in that file, by retailer and period. Covirage does not estimate share; it computes it from the data you give it, and says which file it read.

Can it judge a promotion fairly?

It computes the baseline from the weeks before, the lift during, and the dip after, and nets them. Then it puts net margin against the trade funding for the event. The judgement of whether a 30 percent net lift was worth it is yours; the arithmetic is not left to a model.

What files does a brand need?

Shipments by SKU, retailer and month is enough for the bridge and concentration. Add market data for share, the distribution file for voids, the promotion calendar for lift, and deductions for chargebacks. Each adds a question; none is required to start.

See it on your data

Bring a few thousand rows. The data map opens next, every column mapped once, and the first question is answered in minutes. Free, in your browser, no account.