Demand Planning and Channel Analytics

Aggregate demand forecasts are almost always accurate enough to look fine in a monthly review. The problem lives at the SKU level. A plant producing 120 product variants may show an overall volume attainment of 96% while simultaneously running a stockout on 18 high-velocity SKUs and accumulating 45 days of excess on 22 slow movers. Both problems are invisible in a category-level view, but each has a direct impact on margin, working capital, and customer service levels.

Most Indian manufacturers generate demand forecasts through a combination of sales team inputs, historical shipment averages, and scheme-adjusted uplift estimates. These forecasts are then converted into production plans and raw material procurement schedules. When SKU-level accuracy is poor, the downstream consequences are: production runs adjusted at the last minute to compensate for unplanned demand, expedited raw material procurement at premium prices, overtime shifts to cover shortfalls, and excess finished goods inventory when the forecast overstated demand.

FireAI connects your ERP sales orders and billing data with distributor secondary sales reports to calculate forecast versus actual at the SKU level, not just at the category or geography level. Every SKU gets a forecast accuracy score, a bias indicator (consistently over-forecast or under-forecast), and a mean absolute percentage error (MAPE) tracked over a rolling 12-week window.

What FireAI tracks for demand forecast accuracy:

• SKU-level MAPE: forecast error as a percentage of actual demand, tracked weekly and aggregated by month. SKUs with MAPE above 30% are flagged for demand planning review
• Forecast bias: is a SKU consistently over-forecast (leading to inventory buildup) or under-forecast (leading to stockouts)? Bias is more actionable than error alone because it points to a systematic issue rather than random noise
• Stockout instances attributable to forecast miss: SKUs where a confirmed stockout event occurred within 2 weeks of a significant under-forecast. This connects forecast quality directly to service level impact
• Inventory excess attributable to forecast miss: SKUs where inventory exceeds 30 days of coverage and the forecast over-called demand by more than 25% in the same period
• Forecast contributors: which input data -- sales team input, historical average, or scheme uplift -- has the highest correlation with forecast accuracy for each SKU category? FireAI identifies which input source to trust more by product type
• Week-by-week tracking: for new SKUs or seasonally volatile SKUs, week-level tracking with the ability to update rolling forecasts mid-month based on early actual data

Real example: A Rajkot-based consumer durables manufacturer with 140 active SKUs used FireAI to find that 32 SKUs had MAPE above 35% over a 6-month period. Of these, 14 were consistently under-forecast -- all in the small appliance category where sales team inputs systematically discounted dealer demand signals. Adjusting the forecast weight toward distributor secondary sell-out data for these 14 SKUs reduced MAPE from 38% to 16% over the following quarter and eliminated 4 stockout incidents that had been recurring monthly.

FireAI natural language queries:

• "Which SKUs have forecast accuracy below 70% in the last 8 weeks?"
• "Show me the 10 most over-forecast SKUs this month and their current inventory coverage"
• "Which product categories have the highest forecast bias toward over-forecast?"

New product launches in manufacturing involve significant upfront investment: tooling and die development, pilot production runs, sample distribution, trade channel loading, and launch promotions. Most of this cost is committed before a single unit reaches the end consumer. What happens after initial channel loading -- whether the product actually sells through from distributor to retailer to consumer -- is often invisible until three months have passed and the next ordering cycle reveals whether distributors are reordering or sitting on loaded inventory.

The gap between primary sales (manufacturer to distributor) and secondary sales (distributor to retailer or end customer) is the most important signal in any new product launch. A product that shows strong primary sales in month one but weak secondary sell-through in month two is not gaining traction -- it is accumulating at the distributor level, which will eventually result in returns, scheme pressure from distributors, and production schedule disruption.

FireAI tracks new product launch performance from week one at the SKU level, connecting your ERP shipment data with distributor secondary sales reports, scheme redemption records, and channel feedback to produce a real-time sell-through picture that goes beyond shipment volumes.

What FireAI tracks for new product launch performance:

• Primary vs secondary sales gap: for every launch SKU, what is the ratio of manufacturer-to-distributor shipments versus distributor-to-retailer sales in each week? A widening gap signals inventory loading without demand creation
• Sell-through rate by week: what percentage of the inventory loaded at distributors has been sold to the next channel within each week? Week 4 sell-through rate below 40% is a strong early warning for a launch in trouble
• Stocking reach: how many unique distributors and retailers are stocking the new SKU compared to the plan? Reach tells you whether distribution is actually happening or whether primary sales are concentrated in a few large stockists
• Velocity per retail outlet: for outlets that are stocking the new SKU, what is the average units sold per outlet per week? Low velocity at stocked outlets points to a pricing, placement, or product awareness problem rather than a distribution problem
• Reorder rate: what percentage of distributors who received the first shipment have placed a second order? A low reorder rate within 6 weeks of launch is a critical failure signal
• Scheme redemption tracking: for launch promotions like consumer offers, retailer margins, or distributor incentives, what is the redemption rate? Unredeemed launch schemes indicate schemes are not reaching the intended channel level or are not compelling enough to drive stocking

Real example: A Ludhiana-based agricultural equipment accessory manufacturer launched 4 new SKUs in the Punjab and Haryana markets in February. FireAI flagged that while primary sales were on plan at week 2, secondary sell-through for SKU NP-06 was only 18% by week 4. Investigation showed that the product had been loaded at depot-level stockists who did not service the rural retail outlets where demand existed. A redistribution instruction was issued at week 5, moving inventory from 3 depot stockists to 18 rural distributors. By week 8, sell-through reached 62% and the reorder rate climbed to 44%.

FireAI natural language queries:

• "What is the week-4 sell-through rate for each new SKU launched this quarter?"
• "Which launch SKUs have a primary-to-secondary gap of more than 40%?"
• "Show me the reorder rate for each new product by distributor territory"

Most Indian manufacturers sell through multiple channels simultaneously: direct key accounts, regional stockists, sub-distributors, institutional buyers, and in some cases their own company depots. The revenue contribution of each channel may be visible in billing records, but what is rarely visible is the margin contribution by channel, the working capital intensity by channel, or the trend in channel mix shift and what that shift means for overall profitability.

A manufacturer whose revenue is growing at 12% annually may be experiencing a silent margin compression if the growth is entirely coming from institutional buyers or scheme-heavy stockist channels that carry lower net realization than direct accounts. Revenue stays up; margin comes down. The problem only surfaces in a quarterly P&L review after the damage has accumulated.

Channel mix analysis in FireAI goes beyond revenue split. It connects billing data with scheme spend, trade discounts, freight and logistics costs, payment terms, and collection cycles to produce a true net margin and working capital contribution figure for each channel. Commercial teams can then make pricing, promotion, and channel investment decisions based on what each channel actually contributes to the business rather than what it contributes to topline.

What FireAI tracks for channel mix and contribution analysis:

• Revenue split by channel: what percentage of total billing comes from each channel in each month? Tracked as a trend so that mix shifts are visible before they become significant
• Net realization by channel: after scheme deductions, trade discounts, freight-to-site costs, and credit note adjustments, what is the actual revenue per unit by channel? Channels that appear large on gross billing often shrink significantly on net realization
• Contribution margin by channel: net realization minus variable cost of goods gives contribution per unit per channel. Channels with high volume but low contribution margin are candidates for pricing action or volume reduction
• Payment terms and collection cycle by channel: institutional buyers and government accounts may demand 60 to 90 day credit, which locks working capital for months. The cost of that credit reduces effective margin further
• Scheme spend intensity by channel: which channels require the highest scheme and trade promotion spend per rupee of revenue? High scheme intensity relative to secondary sell-through suggests scheme spend is subsidizing distributor margins rather than driving consumer demand
• Channel concentration risk: is revenue becoming more concentrated in fewer channels? A single stockist or institution representing more than 15% of billing creates collection and continuity risk

Real example: A Coimbatore-based industrial pump manufacturer found through FireAI channel mix analysis that institutional buyers, which represented 34% of revenue, had an effective contribution margin of 8.4% after project-specific freight, extended credit costs, and escalation clauses were accounted for. The company's overall target margin was 14.2%. Regional stockist channels, at 41% of revenue, were delivering 18.6% contribution margin with 30-day payment cycles. The data prompted a deliberate shift in the sales team's institutional bid strategy, prioritizing only projects above a minimum margin threshold. Over two quarters, institutional share dropped from 34% to 24% while overall margin improved by 2.1 percentage points.

FireAI natural language queries:

• "What is the contribution margin by channel this quarter?"
• "Which channels have seen the largest revenue share shift in the last 6 months?"
• "Show me scheme spend as a percentage of net revenue by channel"

Distributor stocking health is one of the most consequential and least-tracked metrics in Indian manufacturing. The standard practice of measuring sell-in from the manufacturer to the distributor tells you very little about whether the market is absorbing the product. A distributor who received 45 days of stock in March as part of a quarter-end scheme push may not order again in April or May -- not because demand has fallen, but because the working capital tied up in slow-moving loaded inventory leaves no headroom to replenish fast movers. The result is stockouts at the retail level on the products customers actually want, while the distributor warehouse is full of products that are not moving.

Replenishment analytics in FireAI tracks the actual inventory position at the distributor level -- not just what has been shipped to distributors, but what distributors are actually carrying and what is moving through to the next channel. By connecting ERP shipment records with distributor secondary sales reports and stock-on-hand data from distributor point-of-sale or manual reports, FireAI produces a real-time picture of stocking health across your distributor network.

What FireAI tracks for distributor stocking and replenishment:

• Days of coverage by SKU by distributor: how many days of secondary sales demand does the current distributor inventory represent for each SKU? Norms typically range from 14 to 28 days. SKUs above 45 days coverage have excess stock that is blocking working capital and reducing reorder urgency
• Fast-mover stockout risk: which high-velocity SKUs are at less than 7 days of coverage at one or more distributors? These require urgent replenishment instructions before a stockout occurs at the retail shelf
• Slow-mover accumulation: which SKUs are sitting above 45 days coverage at multiple distributors? Widespread slow-mover accumulation is a signal of an over-enthusiastic primary sales push that did not align with real market demand
• Replenishment frequency: how often is each distributor placing a replenishment order relative to the expected cycle? Distributors whose order frequency has dropped below their normal cycle are often carrying excess inventory or experiencing cash flow constraints
• Scheme-driven vs demand-driven stocking: for periods following a primary sales scheme or quarter-end push, does secondary sell-through recover within 3 to 4 weeks, or does the inflated inventory persist? Persistent post-scheme inventory indicates scheme-driven loading without underlying demand
• Working capital coverage by distributor: estimated total inventory value at each distributor relative to their typical working capital and credit limit. Distributors near credit utilization limits are unable to place new orders even when fast movers are running low

FireAI natural language queries:

• "Which distributors are carrying more than 40 days of slow-mover stock right now?"
• "Show me fast-mover SKUs at risk of stockout across the South zone distributors"
• "Which distributors have not placed a replenishment order in the last 21 days?"