Demand Forecasting in Tough Times
The key to any Sales & Operations Planning cycle is to start with unconstrained demand forecast which is mostly generated using Statistical Forecasting Methods based on past history data. This works well when you have good historical data – mainly based on Deliveries against Sales Orders or otherwise. Demand Planning moduel of SCM APO offers standard Statistical Forecasting functionalities with Outlier Correction, Adjustments etc. After Statistical Forecast is generated its possible to impose Promotions and modelling of the effect on Sales cannibilised due to Promotion of complimentary products.
In these tough times when pretty much everyone is reeling under the effects of recession (let’s not get in the debate if technically recession is still there or its over) its important to modify the demand signal suitably. Otherwise you end up doing Supply Planning for product quantities which anyway are not sold ending up blocking more cash. Most common is to generate Statistical Forecast and then apply a negative adjustment to reduce the Demand signal. This is fine for the immediate future but how does it get captured for future Statistical Forecasting. If one does a bit of future thinking the challenge then will be how to uplift / take out the impact of recession from Sales History so as to get back to improved Demand Forecast when recession is over. If the Demand Forecast is not uplifted in time there will shortages (Supply Planning was not done for increased anticipated Demand) which may cause irrepairable damage for the company.
As a functional consultant I got similar queries from other projects and interestingly enough this week from my own client. So the hunt began – what is the best approach in these tough times. The solution needs to be something simple and based on standard available features – preferably no major configuration change not to mention any custom enhancemenet (hopefully by the time the change request gets approved and budget allocated for development, recession gets over).
It may sound crazy but using Negative Promotions can be a good solution approach. This is how it will work – setup negative promotions (absolute or percentage) so that the Corrected Forecast is brought down to a realistic level given the business climate. So much for the future demand forecasts. Now as you are coming to the end of recession first make sure as the promotions go in past to save them. At the same time the Sales History collected will be lower in quantity dues to recession effects. Based on this reduced Sales Histroy the Statistical Forecast will show a dip. But you need to take out the impact of recession from the Sales History. This is where the past negative promotions captured in the system will come handy. All you need to do is tweak around in the Promotion sub-section of the Univariate Forecast Profile to enable Change Vals against Promotion keyfigure to correct the Sales History by deducting the promotions (in case of negative promotions adds or uplifts the Sales History).
Once this is done, ensure past Promotions are saved correctly. Check in Interactive Demand Planning as shown below. Note the Sales History / Adjusted Sales History and negative Promotions data in the keyfigure.
Now if Statistical Forecast is executed considering Sales History or in this case Adjusted Sales History (depending on the Read Historical Data in Forecast Profile), the past Promotions are used to uplift Corrected History that is used for Forecast computation. Thus statistical forecast generated is based on past history data normalised for recession trends captured as negative promotions.
The advantage of this process:
First and foremost – this is all standard SAP stuff requiring no configuration only manually setting Forecast Profiles and Promotions.
Secondly – You do not know how long the recession is lasting or going to last. Its possible to continue with one Promotion or create new Promotions as the recession continues thereby giving flexibility.
Thirdly – for a business user this is not very technical but actually makes sense. Interestingly the same process can be applied on products that have sales cannibilised owing to Promotion to complimentary products. Its also possible to record explanatory notes in the Promotion and also report on the promotions in system.