SLIDE 1
Agenda
- Overview
- Methodology
- Conclusion
May 22-23, 2013 MIT SCM ResearchFest 2
Assembling the Crystal Ball: Using Demand Signal Repository to - - PowerPoint PPT Presentation
Assembling the Crystal Ball: Using Demand Signal Repository to - - PowerPoint PPT Presentation
Assembling the Crystal Ball: Using Demand Signal Repository to Forecast Demand Authors: Ahmed Rashad & Santiago Spraggon Advisor: Shardul Phadnis Sponsor: Niagara Bottling LLC. MIT SCM ResearchFest May 22-23, 2013 Agenda Overview
SLIDE 2
SLIDE 3
Agenda
- Overview
- Methodology
- Conclusion
May 22-23, 2013 MIT SCM ResearchFest 3
SLIDE 4
- Demand forecasting technique
- Using external Signals
- Aggregated in a single Repository
What is Demand Signal Repository (DSR)?
May 22-23, 2013 MIT SCM ResearchFest 4
External Signals Repository (Database)
SLIDE 5
When to use DSR?
- 1. What are we forecasting?
- 2. What data is available?
- 3. What stage in the product lifecycle?
- 4. Is the investment worth it?
May 22-23, 2013 MIT SCM ResearchFest 5
SLIDE 6
When to use DSR?
May 22-23, 2013 MIT SCM ResearchFest 6
Trends and Patterns
Time Series
Special Events
Qualitative
Special Events + Trends and Patterns
DSR
- Depends on what are we forecasting
Base Demand Trend Seasonality Unexplained Time Demand
SLIDE 7
When to use DSR?
- Depends what data is available
May 22-23, 2013 MIT SCM ResearchFest 7
Sufficient History
Time Series
Little or No History
Qualitative
Sufficient History + External Data
DSR
SLIDE 8
- Depends on stage in the product lifecycle
When to use DSR?
May 22-23, 2013 MIT SCM ResearchFest 8
Time
Sales Qualitative Time-Series Causal Causal Qualitative
Introduction
Growth Maturity Decline
SLIDE 9
When to use DSR?
- Depends on the investment
May 22-23, 2013 MIT SCM ResearchFest 9
Forecast Accuracy Costs Total System Cost Cost of Inaccuracy Cost of Forecasting Target Area
SLIDE 10
How can we develop a Demand Signal Repository (DSR) to better predict demand?
May 22-23, 2013 MIT SCM ResearchFest 10
SLIDE 11
Agenda
- Overview
- Methodology
- Conclusion
May 22-23, 2013 MIT SCM ResearchFest 11
SLIDE 12
Method Used
May 22-23, 2013 MIT SCM ResearchFest 12
Initiation
- Planning
- Literature Review
- Interviews
- Requirements
Data Management
- Collection
- Validation
Modeling
- Initial Models
- Analysis
- New Models
SLIDE 13
Modeling
May 22-23, 2013 MIT SCM ResearchFest 13
Product
All cases All liters Category SKU
Customer
All Niagara Top 12 Top 3
Geography
All Niagara Region State City 3-Digit Zip code
Time
Annual Quarterly Monthly Weekly Daily Growth Seasonality Wholesale Price Merchandizing Retail Price Natural Disasters Weekly Cycles Buying Patterns Temperature Food Stamps
- 240+ Models
- 60%+ Customer – State - Daily
- 85%+ Customer – State - Weekly
Dependent Variables Independent Variables
Liters per Customer, in a State, per day or week
SLIDE 14
Agenda
- Overview
- Methodology
- Conclusion
May 22-23, 2013 MIT SCM ResearchFest 14
SLIDE 15
Key Findings
- Most Significant:
- Ordering patterns & POS quantity
- Seasonality
- POS revenue (proxy for price)
- Least Significant:
- Temperature
- POS quantity and revenue
May 22-23, 2013 MIT SCM ResearchFest 15
SLIDE 16
Challenges and Caveats
- Accuracy vs. Practicality
- Recording Data
- Retailer Policies
- How much Technology?
May 22-23, 2013 MIT SCM ResearchFest 16
SLIDE 17
Conclusion
- DSR could Significantly increase forecast accuracy
(60%-85%)
- Accurate models are good, Simple models are
better (>5 Factors)
- Perceptions can be misleading (Temperature)
May 22-23, 2013 MIT SCM ResearchFest 17
SLIDE 18