Supply Chain Analytics In FMCG

FMCG demand planning in India typically runs on Excel models built by a central planning team. These models update monthly, operate at the category level, and miss SKU-region combinations where demand patterns diverge significantly from category averages. A shampoo SKU may grow 12% in Tamil Nadu while declining 8% in Rajasthan during the same quarter, but the category forecast shows a flat 3% growth nationally.

FireAI builds SKU-level demand forecasts by region automatically from your DMS secondary sales data and Tally primary dispatch records. The system identifies seasonality, trend shifts, and promotional lift at the SKU-region intersection and generates weekly rolling forecasts with confidence intervals.

What FireAI delivers:

• Weekly SKU-level demand forecasts by territory, state, and zone
• Confidence intervals so procurement teams can plan for best-case and worst-case scenarios
• Promotional lift quantification: how much incremental demand each trade scheme generated historically
• Forecast accuracy tracking (WMAPE) updated every week so the planning team knows when to trust the model and when to override
• Automatic alerts when actual demand deviates more than 20% from forecast in either direction

This shifts demand planning from a monthly category exercise to a weekly SKU-region process, reducing both stockouts and overstock across the distribution network.

FMCG companies in India write off 2-5% of revenue annually to expired or near-expiry stock. The root cause is rarely demand failure alone. It is a visibility problem: batches with 60-90 days of remaining shelf life sit in slow-moving depots while faster outlets in adjacent territories could have sold them. By the time the depot manager flags the issue, the return window has closed and the stock becomes a write-off.

FireAI connects batch-level inventory data from Tally (or WMS) with secondary sales velocity from DMS to calculate days-to-expiry vs days-to-sell for every batch in every location. When days-to-expiry falls below the estimated sell-through window, FireAI raises an alert with a recommended action: liquidate through a specific outlet cluster, transfer to a faster-moving depot, or run a targeted trade scheme.

What FireAI monitors:

• Batch-level shelf life remaining across all depots and C&F points
• Current sell-through velocity per SKU per location (units/week)
• Days-to-sell estimate: current stock / weekly run rate
• Expiry risk score: batches where days-to-sell exceeds days-to-expiry
• Recommended redistribution: which nearby depots have higher velocity for the same SKU
• Financial exposure: total value (at MRP and at cost) of at-risk batches

This converts expiry management from a reactive quarterly exercise into a continuous automated process that catches risk 60-90 days before write-off.

Most FMCG companies in India operate with 15-30% more SKUs than they need. Long-tail SKUs that contribute less than 0.5% of revenue individually often consume disproportionate warehouse space, procurement bandwidth, and working capital. But rationalising SKUs without data is risky: a slow-moving national SKU may be a top seller in 2-3 specific territories, and removing it would hurt those markets.

FireAI performs SKU rationalisation analysis by combining revenue contribution, margin contribution, velocity, geographic concentration, and cannibalisation patterns into a single rationalisation score. The output is not a binary keep/kill list but a tiered recommendation: retain, retain regionally, reduce variants, or phase out.

What FireAI analyses:

• Revenue and margin contribution by SKU at national, zonal, and territory level
• Velocity ranking: units sold per outlet per month, segmented by channel and region
• Pareto analysis: how many SKUs make up 80% and 95% of revenue
• Cannibalisation detection: SKU pairs where adding one consistently reduces sales of the other
• Working capital tied up in slow-moving SKUs: stock days and carrying cost
• Regional dependency: SKUs that rank in the bottom 10% nationally but top 20% in specific territories

This gives the category team a defensible, data-backed rationalisation plan instead of gut-feel decisions that risk alienating regional distributors.

FMCG demand in India is deeply seasonal. Floor cleaners spike 40-60% before Diwali. Beverage sales double in April-June. Soap and personal wash surge during monsoon. Yet most FMCG companies plan for seasonality using last year's numbers adjusted by a flat growth percentage, ignoring that the magnitude and timing of seasonal surges vary by region, channel, and SKU within the same category.

FireAI builds seasonal demand models that decompose historical sales into trend, seasonal, and residual components at the SKU-region level. The system learns not just that demand rises before Diwali, but by how much, how many weeks before the festival the ramp begins, and which SKUs see the earliest uplift. This allows procurement and production teams to phase inventory build-up rather than placing one large order that arrives too early (tying up capital) or too late (causing stockouts during peak).

What FireAI models:

• Festival and holiday demand patterns by SKU, region, and channel
• Lead time between procurement trigger and peak demand week
• Regional variation: Navratri-driven surge in Gujarat vs Onam-driven surge in Kerala for the same category
• Weather-correlated demand shifts: monsoon impact on personal care, summer impact on beverages
• Post-surge demand drop modelling to prevent overstocking after the seasonal peak

The output is a phased procurement calendar that tells supply chain teams exactly when to place orders, in what quantities, and to which depots, so inventory peaks align with demand peaks rather than arriving 3 weeks early or 2 weeks late.