Forecasting and customer order management

Question 1

WHat’s bullwhip effect?

Answer 1

The further down in the supply chain, the heavier the fluctuations get.

Question 2

FActors influecing the forecast?

Answer 2

• Forecast data
• The length of the forecast period – e.g. week, month, year?
• Forecast frequency – how often we redo our forecast?
• Forecast horizon – how many periods ahead we are looking?
• Aggregation level – product, product group?
• Forecast units – pieces, metres, litres, kilogram, monetary unit, etc?

Question 3

sales management approach?

Answer 3

• Management personnel gather in meeting(s) to forecast demand
• Considerations to sales statistics
• Forecast breakdown into regions
• Quick way to produce forecasts
• Consensus achieved among (participating) management
• Risk of subjective assessments
• Management decides – rather than the individuals closest to the marke

Question 4

Grassroot approach?

Answer 4

• Initial assessments made by salespeople and other personnel in direct contact with
the market
• The initial forecasts are collected and processed centrally
• A common forecast is created for the whole company
• Forecasts produced by those with the most knowledge of the market
• Responsibility to realise forecasts falls upon those who made them
• Forecast breakdown into regions not necessary
• Takes more time than the sales management approach
• Risk of subjective assessments

Question 5

Pyramidal approach?

Answer 5

• Both management and grass root participation
• Bottom up – top down
• If the grassroot forecasts deviate from the management forecasts, grassroot
forecasts are proportionally adjusted
– The original proportion between e.g. regions or product areas remain
– The total equals management total
• Sales staff not tied to its own forecasts

Question 6

How do we use moving average to forecast?

Answer 6

(D(1)+…..+D(n))/n

Question 7

Formula for exponential smoothing?

Answer 7

F(t+1)=F(t)(1-a)+D(t)a

Question 8

How does focus forecasting work?

Answer 8

• Simple logic: If it worked before, it will probably work again
• For every forecast occasion tests are made by running simulations
to find which method that would have worked best in the previous
period/s, where the outcome is known.
• The best working method is selected to generate the forecast for
the next period
• Assumes access to large computer capacity