Demand Planning -Pharmaceutical Industry

The Indian pharmaceuticals market is the 3rd largest in terms of volume and 13th largest in terms of value, as per a report by equity master. India is the largest provider of generic drugs globally with the Indian generics accounting for 20 percent of global exports in terms of volume.

According to India Ratings, a Fitch company, the Indian pharmaceutical industry is estimated to grow at 20% CAGR over the next five years. Presently the market size of the pharmaceutical industry in India stands at US$ 20 billion.

The client, being a large listed Indian pharmaceutical company with revenues exceeding US$ 2 billion, wanted to move from legacy practice of arbitrary estimation using excel sheets, to a scientific basis for Demand Planning & Forecasting (DP&F).

Business Need

• Prior to IBM Cognos TM1 implementation, DP&F was done for only 20% (~300) of total SKUs, at a National-level, basis immediately preceding 3 months’ sales salience. This process delivered a forecast accuracy of ~73%.

• Multiple hierarchies existent in the client’s sales structure in the form of Area, Territory, CFA, Region, Division meant having to choose 1 level of forecasting that gives the best results.

• The outlook to be provided to the supply chain was arbitrarily based on Demand Planner’s vision, with few inputs from Sales Managers &Division Heads. This was a time-consuming affair that took up to 3 weeks for data collation and coordination.

• No involvement of Regional Sales Managers (RSMs) in the demand planning process, despite being resources in direct contact with customers/stockists and possessing ground-level intelligence of events.

• Redundancy of multiple years’ historical data available in SAP as Demand Planning Team (DPT) exerted dependency on merely 3 months’ historical data.

• Number of future months to be considered for forecasting.

The Solution

• IBM Cognos TM1 combined with IBM SPSS Statistics was suggested as the best replacement for the legacy system. With the established success of connectivity with SAP BW, data integration wasn’t much of a task.

• 5 years’ gross sales data (normal sales + free schemes) was suggested as the most relevant measure to perform statistical forecasting as it reflects actual sales including schemes and does not factor in returns.

• RSMs (Regional Sales Managers) were selected as appropriate-level for intelligence input due to considerations:

• Best resource having ground-level intelligence.

• Minimal impact on daily work.

• Less complexity involved in tool roll-out.

• A Pareto Analysis was done of all active products to exercise greater forecasting focus on those medicines with high sales salience and low coefficient of demand variation.

• The sales structure being dynamic, the sales hierarchies were designed so as to be recreated on any updates in SAP masters for every monthly forecast cycle.

• Uploading of 5 years’ historical data was designed in a manner so as to upload 1 month’s data at a time, which optimized bandwidth restrictions on client network and avoided process failure due to network drop.

• Considering supply chain lead time for different sourcing channels and also the fact that near-term forecasts are more reliable, it was suggested that forecasting be carried out for 18 future months with the main emphasis on and consensus for mainly 4 future months.

The Benefits

• DP&F cycle time drastically reduced from 1-2 weeks to barely ~3 hours.

• Increase in forecast accuracy from ~71-73 % at a National-level for 20% of total SUs, to 83% forecast accuracy for 100% SKUs, up to 1CFA-Region combination-level for the horizon of 4 future months.

• Historical Data imported into IBM Cognos TM1with 100% accuracy means one version of the truth, resulting in reduced dependency by DPT for generating data via SAP tables.

• Ridding arbitrary forecasting by DPT (earlier excel based) by ensuring there is a scientific base for demand planning. Also, the availability of audit trail for values before and after outlier correction by Demand Planner through forecast file archiving.

• Process reengineering instead of process improvement by introducing a 2FPM system in place of legacy systems that largely excelled dependent.

• DPT empowered with enhanced capability to forecast in a drastically reduced period of time.

• Control Center for carrying out forecast conveniently designed to enable the user to complete forecast on merely clicking step-by-step 7 action buttons.

• One single platform for enabling the collaboration of a consensus forecast with the involvement of a broader base of stakeholders.

• Removal of the possibility of arbitrary demand planning as per DPTs best understanding thus demolishing any case of instability in the demand planning process in case of attrition within the team.