1 of 26
FORECASTING IN REGULATED MARKETS(US & UK)
Kaustubh V. Kokane
Summer Trainee- Supply Chain
Mentor- Mr. Umang Gandhi (GM- Supply Chain)
27th June, 2013Thursday
2 of 26
Flow of Presentation
Forecasting Industry best practices US forecasting
Forecasting process Pros & cons Challenges Suggestions for improvement
UK Forecasting Forecasting process Pros & cons Challenges Suggestions for improvement
3 of 26
Forecasting
Predicting future demand Short and mid-term forecast tactical / operational
planning (1 month to 1 year) Long-term forecast Capacity planning (1 year +)
Critical point of co-operative work between supply chain and marketing teams
Forecast models Time horizon Time-series analysis
4 of 26
Maturity-wise Forecasting
Criteria Matured Drugs Growth Stage Drugs New Drugs
Availability of historic sales data
Ample Short-span sales history
No sales history
Forecasting Extensive quantitative forecasting possible
Short-term forecasting possible
Focus on qualitative forecasting with inputs regarding launch scenario, promotions, etc.
Forecast Accuracy Expected to be more accurate
Accuracy will improve over time
Difficult to forecast with great accuracy
5 of 26
Branded v/s Generics
Branded Drugs Generic Drugs
Indian market dominated by branded drugs
US & EU markets dominated by generic drugs
Drugs prescribed & sold by their trademarks
Drugs prescribed & sold by their generic name (molecular name)
Usually costlier than generics Usually cheaper than branded drugs
R&D expenses involved for manufacturer Drug-discovery R&D expenses not involved
Patent protection for the manufacturer Generic players come into picture only after a molecular patent expires
Major companies- Pfizer, GSK, Novartis, J&J, Roche, Sanofi
Major companies- Teva, Mylan, Sandoz, Greenstone, Hospira, Dr. Reddy’s
6 of 26
Replenishment v/s Forecast Models
Criteria Replenishment Model Forecast Model
Production Demand-driven Forecast-driven
Inventory Decreased retailers’ & manufacturers’ inventory
Inability to meet changing demand patterns & risk of obsolation
Flexibility More responsive Not flexible enough to handle frequent changes
Information Automated information processing through VMI and collaborative planning
Risk of inaccurate information
Lead time Decreased because of availability of inventory at each level
Predictable lead time
Industry example
GSK, Teva Pharmaceuticals, Dr. Reddy’s (domestic)
All generic pharma exporters based in India, Ranbaxy (domestic)
7 of 26
Effects of Ineffective Demand Planning
Higher production cost Overstocking or stock-outs Inefficient logistics Dissatisfied customers- losing to competition FDA cracking up on expired drugs
8 of 26
INDUSTRY BEST PRACTICES
9 of 26
Data used for Forecasting
Primary Data
Point-of-sale data not readily available
Proprietary patient data (through third-party services such as e-talk’)
Secondary Data
Purchase data from distributors (fee-for-service)
Vendor-managed inventory examples:
(US)
(Denmark) Channel data (ex. Retailers’ IMS
access to pharma co’s- - partnership) Utilize sales force (feedback from
retailers)
10 of 26
Pfizer (Australia)
Integrated demand forecast review process in business planning
11 of 26
Teva Pharmaceuticals, UK
Was relying on stand-alone Excel sheets for forecasting large no. of SKU’s (till 2009)
Laborious and time-consuming process Implemented customized demand management
tool- RefleX to forecast & plan demand Forecast accuracy improved from 65% to 80% Seamless business integration thereafter
12 of 26
FORECASTING @ WOCKHARDT
13 of 26
Forecasting Process- US
Manual Adjustments(Supply Chain team)
Customer Forecasts & Inventory Data
(EDI)
John Galt SolutionsForecasting Software
Sales & Marketing Inputs
Demand Forecast
Sales History & Trends
Measuring Forecast Accuracy
Highlights: 18-months rolling forecast Forecast accuracy (MAPE method): 75% (A-class), 66% (overall) Forecast accuracy measured by MAPE, quantitative bias
14 of 26
Product Group-wise Forecast Error (US)
Average Forecast Error:
For matured products: 34%
For non-matured products: 70%
Note: Forecast data from January-13 to April-13 was analyzed for the above results
15 of 26
Strong quantitative fundamentals & future plans
Long horizon for forecasting- capacity planning
Industry-standard MAPE method of measuring forecast accuracy, can be benchmarked
Working capital management
Poor visibility from customer-end (no customer collaboration)
All marketing inputs may not reflect in final forecast (as supply chain team takes the final call)
Pros Cons
Forecasting @ Wockhardt US
16 of 26
Challenges- US
Controllable: Lack of visibility (customer side) Penalties related to tendering
Non-controllable: Demand volatility Sales concentration- top 3 wholesalers Market is very sensitive to short-supplies
17 of 26
Suggestions
‘Bias’-based forecast accuracy monitoring Consistently lower or higher forecast than actual sales In April 2013 (forecasted over 4 months):
30% SKU’s – Under-forecasted 18% SKU’s – Over-forecasted
Causes: Undetected patterns “Beat the numbers” approach
Weighted bias / size-based bias
18 of 26
Suggestions
Shift from MAPE to WMAPE (Weighed Mean Absolute Percentage Error) Yields more meaningful analysis of forecast accuracy
19 of 26
Suggestions
Consensus forecasting: Particularly for non-matured drugs (high uncertainty) Concurrent forecasting team with SC and marketing
executives (S&OP) Review supply and demand requirements frequently
20 of 26
Forecasting Process- UK
Sales Forecast (Marketing team)
Supply Chain team(Past trends, stock
levels & other inputs)
Manual Adjustments(Supply Chain team)
Demand ForecastMeasuring Forecast Accuracy
Highlights: 12-months rolling forecast SKU category-wise forecast accuracy analysis
21 of 26
Strong marketing team inputs
Long horizon for forecasting- capacity planning
Poor visibility from customer-end (no customer collaboration or VMI)
No dedicated forecasting software solution in place
Forecast accuracy cannot be benchmarked with current metric
Forecasting @ Wockhardt UK
Pros Cons
22 of 26
Challenges- UK
Controllable: Customer collaboration Service level management (90-95%)
Non-controllable: Demand volatility Market is very sensitive to short-supplies
23 of 26
Suggestions
EDI of inventory and sales data
Established standard in US & EU Modeling & analyzing the inventory data Suitable alternative to VMI for mid-level producers Better visibility of customer demand trends Better tackling of bullwhip/whiplash effect
Top customers Wockhardt
Forecast & inventory data
Better serviceability
24 of 26
Suggestions
Quarterly benchmarking against competitors for forecast accuracy and market share Shifting to industry-standard metric of measuring
forecast accuracy (MAPE / WMAPE)
Particular Forecast Accuracy
US Generic Marketers ~70%
Wockhardt US 60-70%
Teva Pharmaceuticals UK 85%
25 of 26
Suggestions
Validate forecasts for established drugs: Scientific validation in statgraphics Remove bias and better selection of forecasting model
Quantitative forecasting software: Optimum use of available data; data mining Useful for established products
26 of 26
Kaustubh V. KokaneSummer Trainee- Supply ChainMentor- Mr. Umang Gandhi (GM- Supply Chain)
Institute- Prin. L. N. Welingkar Institute of Management, Mumbai
Thank You !
Top Related