Volume forecasting from concept tests Volume forecasting predicts take-up of new products/services Forecasts are based on likelihood of purchase and value for money responses to concepts/products, coupled with information concerning distribution and awareness The technique is centred around the classic trial/repeat model, with extra stages built in to allow comprehensive assessment of sensitivity to variation in key input parameters The trial/repeat model is implemented in the form of a 2-state Markov Chain The sensitivity analysis is conducted via a Monte Carlo simulation, coupled with resampling to enable accurate determination of statistical confidence intervals for the forecasts Volume forecasting from concept tests (cont.) Output example, showing minimum, maximum and most likely range estimates 1400 1315 1200 Estimated Tonnes in Year 1 80% confidence interval 1000 938 800 767 754 637 600 548 543 444 400 399 358 348 288 260 281 246 207 200 0 Option 1 Option 2 Option 3 Option 4 Option 5 Option 6 Option 7 Note: Volumes for different options cannot be added to yield a total volume of sales Option 8