Fully integrated AX module Demand forecasts Sales orders Transfer orders Purchase orders Production orders Master Planning Demand Forecasting.

Fully integrated AX module

Demandforecasts

Sales orders

Transferorders

Purchaseorders

Productionorders

Master Planning

DemandForecasting

Let’s look at an example

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 240

50

100

150

200

250

300

350

400

450

500

Demand History

Let’s look at an example

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 240

50

100

150

200

250

300

350

400

450

500

Demand History & ForecastDemand HistoryForecast

Now let’s compare the results

  Forecasting No Forecasting

Average Forecast Error % 35% -

Now let’s compare the results

  Forecasting No Forecasting

Average Forecast Error % 35% -

Average Safety Stock 144 144

Average Stock Level 261 295

These results compare the replenishment of stocks over an 18 month period for the item on the previous screen, assuming a 3 month replenishment lead time and a safety stock factor of 1.65 (95% service level)

Now let’s compare the results

  Forecasting No Forecasting

Average Forecast Error % 35% -

Average Safety Stock 144 144

Average Stock Level 261 295

Service Level % 100.0% 98.5%

Now let’s compare the results

  Forecasting No Forecasting

Average Forecast Error % 35% -

Average Safety Stock 144 144

Average Stock Level 261 295

Service Level % 100.0% 98.5%

If Safety Stocks are calculated based on forecast accuracy, not demand variability, even greater savings can be achieved. In this case a 17% reduction in average stock levels are produced, while still achieving a 100% service level

Major Points

• Safety stocks should be calculated, not guessed- It’s easy to inflate safety stocks to achieve service levels, but there’s a cost

• Lead times are a major factor- Reducing lead times reduces risk of stock-outs (and reliance on forecast

accuracy)

• Forecasting can be simple- Your existing customers have the data already- It can be implemented quickly & easily

The Forecasting Cycle

Generate demand history

Generate base forecast

Copy to final forecast

Generate forecast

allocations

Review forecasts

Final forecasts may contain additionalitemdimensions(depends on each item’s coveragerequirements)

Demand history is consolidated by SKU, customer groups & forecast periods. A SKU contains all item dimensions from demand transactions

Forecasts only contain the item dimensions specified in each item’s Input Dimension Group

Allocations are used to distribute final forecasts over additional item dimensions (if required)

Sales orders or sales invoices

Project item trans-actions

Inventory BOM Jnl lines, Production BOM lines

Key features

• Forecasts can be calculated automatically or entered manually

• Forecasts can be changed by quantity and/or value and at multiple levels

• Graphical view for comparison of forecasts and demand history

• Adjust forecasts either online, or offline using Excel

• Review forecasts within companies, or summarized over multiple companies

• Built within Dynamics AX – no interfacing required!

Demand Forecasting R6.0

• Flexible selection of transactions when generating demand history

• Multi-level forecasting review

• Forecasts can now be reviewed/changed by value

• Regenerate forecasts for selected item groups only

• Generate forecasts from any date in the past

• Weekly periods can now start on the same weekday

• User-defined number of periods for calculating average sales price

• Improvements to Item Supersessions

Item Supersessions

Demand history for a new item can now be built before the Supersession End Date, and forecasts can be delayed to start at a later date

Start Date

End Date

Forecast Start

Demand copy

New item demand

Old item demand

New item forecast

Demand Forecasting R6.0

• Forecast Model Copy (easy snapshotting of forecasts)

• New Forecast Allocations replace use of Item Allocations Keys

• Allocations corrected for different units of measure

• Budget periods fixed (based on budget copy date)

• Average sales price now calculated using only demand history records with the same item dimensions as the forecasts

• Forecast Groups no longer attached directly to customers (removal of mods to standard AX tables)

Items are grouped for ease of maintenance

Customers are grouped to consolidate demand and improve forecast accuracy

Forecast at any dimension level

Collaborative Forecasting

• Forecast Sales Groups specify who manages forecasts for each customer

• Forecasts can be adjusted online, or offline using Excel

Abnormal Demand

• Ability to flag abnormal demand– For example an abnormally large order is received from a

customer to stock stores for a sale– You want to adjust the demand history to prevent future

forecasts being inflated– Manual adjustments are always held separately from

actual demand history

Promotions

• Ability to enter the likely impact of promotions for forward planning

• A promotion may drive up sales in a period, but reduce sales in a subsequent period

Item Supersessions

• New product– Automatically copy history from a similar or superseded

item

• Old product– Manually determine when the

run-out should occur– Disable item from being forecast– Copy history to a new item

Forecasting Formulae

• Same as last year’s demand +/- n%• n-month moving average• n-month moving average with trend• Average of same months last 2 years• ‘Best fit’ formula selection using advanced

forecasting algorithms

Advanced Forecasting Formulae

• Simple Moving Averages• Discrete Data Models• Croston’s Intermittent Demand Model• 9 Exponential Smoothing Models• Univariate Box-Jenkins Model• Event Models• 4 Curve Fitting Models• Dynamic Multiple Regression

Other Features

• Forecast periods– Forecasts are calculated by week, month or user-definable periods– Monthly forecasts can be automatically split into weekly forecasts (or

even further if required) using period keys

• Forecast accuracy– Multiple forecast models enable a forecast to be frozen and compared

to actual demand over the forecast period

• Fully multi-lingual capable– English language labels are shipped with the product– Other language labels to be developed by resellers

Demand Forecasting Roadmap

• Additional item group

• Multi-level forecast generation

• Automatic detection of demand history ‘outliers’

• Improved forecast exception tracking

• Safety stock calculations

• User-definable forecasting formulas

• Extensions to notes to allow for changes at multiple levels

• Audit recording of changes and who made them

Demand Forecasting R6.1

• Ability to select individual Forecast Pro formulas

• Automatic detection of demand history ‘outliers’

• Improved forecast exception tracking