During my career I have worked as both a Demand and Supply planner, working across FMCG, Fresh Produce and Cosmetics industries. When it comes to creating baseline forecasts for New Product Introduction (NPI) or phasing out forecasts for End of Life (EoL), planners find it challenging to collect relevant information from various stake holders and spend more time crunching data sets than creating meaningful forecasts.
This blog will discuss the features of SAP's Integrated Business Planning (IBP) Product lifecycle application within the Demand module and how it will help planners to overcome these challenges and help create a more accurate base line forecast. Enjoy the read!
Generating New Product Forecasts
Forecasting NPI is difficult due to a lack of sales history and relevant information, so normally in such situations planners use the history of similar products or product groups to generate the baseline forecast (which could be statistically driven).
I would categorise NPI into three categories
- Upgrade or Revision: it is a 1:1 replacement for the existing product, use 100% of old product sale history. Although it may be 225g down size to 200g, the shopper still buys this as the e.g. daily margarine.
- Product Extension: new products are launched to expand the portfolio within an existing category, e.g. a new Lemon & Lime shampoo.
- Innovation: a brand new category of product is launched, where the organisation must complete market surveys or similar activities to come up with the estimated sales volumes.
The SAP IBP Product Lifecycle application provides options to select and use history from one or multiple products by giving weighting to each product as shown in the below illustration. The weighting can be more than 100% on one reference product.
As you can see above, there are two new products AP_F007 & AP_F015.
- AP_F007 will use 100% history of existing product AP_F001.
- AP_F015 will use historical references of 3 different products AP_F001, AP_F002, AP_F003 as per the weights assigned to each of them.
The results below are the output of running a statistical forecast using these settings.
Here, historical sales are captured in key figure ‘Delivered qty’, and statistical forecast is captured in key figure ‘Statistical Forecast Qty’. As we can see in above example, we do not have any historical quantities captured for AP_F007, but the statistical forecast is still created using 100% reference of AP_F001.
Similarly, the above example has new Product AP_F015 which doesn’t have any historical sales, but uses a mixture of the history from AP_F001, AP_F002 and AP_F003 with percentage of 60%, 20% and 20% respectively to create forecast for AP_F015.
‘Manage Lifecycle’ app also provides graphical representation of reference product history which will help planners to visually check history of reference products seasonality, trend etc.
Once new product gains few months of history it can be added as reference product then statistical forecast will consider its own sales history as well as other reference product maintained.
Launch Dimensions
In some situations, new products are planned to have limited launches, such as to a specific customers, customer groups or a region. In such scenarios, the Launch Dimension functionality will help planners to specify the respective launch dimensions and create profiles so statistical forecast is created only for specified launch dimensions.
Let me explain this with an example. We will take new product AP_F007 and plan to launch only to a specific Customer Group called Strategic. In this scenario while setting profile for new product, Customer Group will become Launch Dimension and it will allow to drill down into existing customer groups to select any specific one, in this example it will be Strategic customer group.
When the Statistical algorithm runs it will consider the above setting and take only strategic customer group history of AP_F001 as this is reference product and generate future statistical forecast for AP_F007.
As you can see above statistical forecast of new product AP_F007 is generated only of Strategic Group and for reference product AP_F001 it generated both a group as it is ongoing product for both existing groups.
New Product Transition Period
Markets take time to accept any new product whenever it is launched and this period is known as the transition period. If a product is replacing another, during that period the forecasts need to gradually increase and decrease accordingly. Through this application planners can choose length of transition periods, and how the forecast should be deployed during that period – month 1 = 20%, month 2 = 30% etc.
Let’s take example of Product ‘AP_F007’, as seen in the below picture, the phase in start date is “1st October” which means that the product will be launched on 1st of October, and phase-in end date is “31st March”, which means that this product would be fully phased in by 31st of March 2019. So, there is 6 months of transition period.
There is also another setting called Forecast Start date, which is basically advising system that from what date statistical forecast can be generated for this product.
As mentioned earlier, this application also provides the option of managing the transition period forecast like how much of statistical forecast should be used in first month and so on. It provides three basic options in form of curve. (Using below curves one can create their own custom curves)
- Linear
- Sublinear
- Superlinear
By setting start value and End value planner can define the behaviour of the curve during transition period.
For new product AP_F007 I have used Sublinear curve for period of 6 months and forecast should start with 10% of statistical forecast in first month, which is October 2018 and end with 90% in March 2019 and use 100% of Statistical forecast from April.
Statistical algorithm considers above setting and provide only 10% of statistical forecast in October and gradually increase to 90% in month of March 2019 and will be same forecast as AP_F001 from April 2019.
Once the new product gains its own history planner can deactivate this profile then Statistical algorithm will take its own history to generate future statistical forecast.
End of Life Forecasting
We could use same application to create a phase out / end of life forecast for existing products. This is done for existing product by creating a lifecycle profile for the end of life product and maintaining reference product as itself.
Let’s say, AP_F001 which is ongoing product and planned to phase out end of 2019 only for strategic customer group. This can be achieved with below setting
New product and Reference product will be same (as there is no new product);
Rest of features like selecting length of transition period and behaviour of forecast during that period will remain same
In, below setting - AP_F001 is not phasing out for Large Customer group hence Phase out dates are not maintained and for Strategic customer Group it will be phased out from 1st of July 2019 and exist market on 31st of Dec 2019.
Used Sublinear approach during this phasing out period forecasting which is in in July 2019 use 90% of statistical forecast and gradually reduce to .01% by end of December 2019 and no forecast from 2020.
Forecasting algorithm considers above setting and gives statistical forecast for Large customer group however, for strategic customer group it will start reducing from July 2019 and ends on December 2019 as per above setting.
Conclusion
The SAP IBP Product Lifecycle application gives more flexibility to planners and help in developing more meaningful baseline forecast (statistical) by taking advantage of its computational power. This application takes away all calculation and data crunching task from planner and give more time to add value to generated forecast. To find out more about this and any other functionality within SAP IBP, get in touch with Olivehorse today!
Business Analyst - Olivehorse Consulting