NEW PRODUCT FORECASTING & NAÏVE FORECASTING Session III Dr.M.I.Haque

NEW PRODUCT FORECASTING Using market research Product life cycle Analog forecast Test marketing Product clinics The bass model

Using market research Identification of a demographic profile of the target market Probability of purchase from survey data….Sample of consumers is asked to respond to intent to purchase scale A forecast is developed by combining this probability with information on the size of the target market

Product life cycle

Analog forecast-Forecast of new product is related to information about information of similar products in past Test marketing-Introducing a product to a small part of the total market before doing a full product rollout Product clinics-Mock market in lab---experience and evaluate the product

Type of product affects new product forecasting High tech products have short life cycles than low tech products Fashion products have a steep introductory phase followed by short growth and maturity stages an decline is also very steep

Product diffusion curves CD player……….black & white tv ………….

The bass model for new product forecasting S t = pm + (q-p)*Y t – (q/m)* Y t 2 S t = sales in time period t p= probability of initial purchase at time t m= number of initial purchases over the life cycle q=coefficient of imitation Y t = number of previous buyers at time t

F t = A t – 1 Naive forecast …………….. F t = A t – 1 Jan120 Feb90 Mar100 Apr75 May110 June50 July75 Aug130 Sept110 Oct90 ORDERS MONTHPER MONTH Nov - FORECAST

II model ……….Naive forecasting F t = F t – 1 + p(A t – 1 - A t – 2 ) Ft = forecast for period t At – 1 = actual observation at period t – 1 At – 2 = actual forecast at period t – 2 P is the proportion of change between the periods t – 1 and t – 2 that we choose to include in the forecast

Evaluating forecasts The mean error = ∑ (A t - F t ) / n= ME The mean absolute error = ∑ │A t - F t │ / n = MAE The mean percentage error = ∑ [ (A t - F t )/ A t ] / n = MPE The mean absolute percentage error = ∑ │ (A t - F t )/ A t │ / n = MAPE The mean squared error = ∑ (A t - F t ) 2 / n = MSE The root mean squared = √∑ (A t - F t ) 2 / n = RMSE Theil's U = √∑ (A t - F t ) 2 ÷ √∑ (A t – 1 ) 2

THIEL U For Thiel U value is 1 to 6. lower values are preferred Thiel value of 0 means model forecasts perfectly If U < 1, model forecasts better than the consecutive period no change naïve model If U = 1, model does only as well as the consecutive period no change naïve model If U > 1, model does not forecasts as well as the consecutive period no change naïve model