Forecasting a Chilled Batter Launch with No History and a Minimum Batch Size
Sunrise Foods is launching refrigerated idli and dosa batter in Bhubaneswar, Cuttack and Rourkela. The product has a 30-day shelf life and needs cold-chain distribution. The company has three distributors and an initial list of 45 retailers. There is no company history for this product, but a similar fermented-food launch in another state achieved 12 packs per retailer per day after six months. The production line has a minimum batch of 5,000 packs per week, and below that the cost per pack rises sharply. Retailer margin is 20%, distributor margin is 8%, and the launch promotion budget is Rs 3 lakh. The sales manager must prepare a six-month forecast, set distributor quotas, and decide whether to launch in all three cities at once. A competitor sells a 500g pack at Rs 45.
1. Recommend a forecasting method for the first six months and build a simple forecast. Explain the assumptions and the main sources of error.
Read the worked answer
With no company history, a purely statistical trend forecast is not credible. The better method is a market build-up forecast, supported by a proxy benchmark and then adjusted by judgement. The build-up is: number of active retailers x packs sold per retailer per day x selling days. The proxy benchmark is 12 packs per retailer per day, but that is a mature-state figure and should not be applied at full strength in month one.
A defensible plan is: month 1 at 60% of benchmark with 45 retailers; month 2 at 80% with 50 retailers; months 3 to 6 at 100% with 55 retailers. Using 26 selling days, month 1 is 45 x 12 x 0.60 x 26 = 8,424 packs. Month 2 is 50 x 12 x 0.80 x 26 = 12,480. Months 3 to 6 are 55 x 12 x 1.00 x 26 = 17,160 each, or 68,640 for four months. The six-month total is 89,544 packs, averaging about 14,924 per month, or about 3,731 per week.
The main errors are benchmark transfer (the other state's taste, price and retail density differ), distribution assumptions (45 listed retailers are not 45 productive retailers), and seasonality (festivals and summer holidays affect batter consumption). The forecast should be revised monthly using actual retailer offtake, not distributor billing, because billing can hide stock in the channel. A 10% upside and downside range should be carried into production planning.
2. Set monthly distributor quotas for month 3 using a territory potential index of 50:30:20 and a 10% stretch. Show the calculation and explain why quota is not the same as forecast.
Read the worked answer
The month 3 forecast is 17,160 packs. A 10% stretch gives a quota total of 18,876 packs. Applying the potential index of 50:30:20 gives Bhubaneswar 9,438, Cuttack 5,663 and Rourkela 3,775 packs, allowing for rounding.
But the index alone is not enough. Rourkela has a lower potential index and higher distribution cost, so a purely proportional quota may be demotivating. The manager should check retailer productivity and travel time in each city before finalising. If Rourkela's 20% share reflects fewer retailers rather than lower demand per outlet, the quota may be fair; if it reflects poor cold-chain coverage, the quota should be paired with a distribution-building target.
Quota and forecast serve different purposes. The forecast is the company's best estimate of demand for production, cash and inventory. The quota is a performance standard for the distributor and sales team. A quota is usually set above the forecast to create stretch, but it should be achievable by a well-run territory. Setting quota equal to forecast removes the incentive to grow; setting it 30% above forecast destroys credibility. A 10% stretch is reasonable in launch, provided the distributor is given the retailer activation plan and the promotion budget to support it. Quotas should also be expressed in packs and in retailer coverage, so that distributors cannot meet the number by loading stock.
3. The production minimum is 5,000 packs per week. Compare that with the month 3 forecast and recommend a launch plan.
Read the worked answer
Month 3 forecast of 17,160 packs is about 3,958 packs per week using a four-week month, or about 3,731 using 26 selling days. That is below the 5,000-pack weekly minimum. If Sunrise Foods launches in all three cities at the forecast level, it will either run the line below minimum and absorb high unit cost, or produce 5,000 packs and risk about 1,000 packs of waste each week. With a 30-day shelf life, that waste is a direct write-off, not a financing cost.
The launch should therefore be phased, not spread across all three cities. Start with Bhubaneswar and Cuttack, where the potential index is 80% of the total. If those two cities achieve 80% of the month 3 forecast, that is 13,728 packs, or about 3,432 per week, still below minimum. The company must either raise retailer productivity, add more retailers before launch, or accept a higher initial cost for a short period. A realistic plan is to launch in two cities with 60 retailers, target 15 packs per retailer per day in the first month, and hold Rourkela until weekly offtake crosses 5,000 packs. Alternatively, use a co-packer for Rourkela to avoid a minimum-batch penalty. The decision should be based on waste cost versus delay cost, not on the desire to be present everywhere on day one.