Ananth Raman - PowerPoint PPT Presentation

1 / 31
About This Presentation
Title:

Ananth Raman

Description:

Harvard/Wharton Project: How can recent advances in ... David's Bridal CompUSA Borders Group Ahold. Footstar Office Depot Bulgari Christmas Tree Shops ... – PowerPoint PPT presentation

Number of Views:260
Avg rating:3.0/5.0
Slides: 32
Provided by: Harv182
Category:
Tags: ananth | depot | office | qvc | raman

less

Transcript and Presenter's Notes

Title: Ananth Raman


1
Forecasting and Inventory Planning in the Supply
Chain
  • Ananth Raman
  • UPS Foundation Professor of Business Logistics,
    Harvard Business School

2
Harvard/Wharton Project How can recent advances
in Information Technology improve the way
retailers forecast demand and plan supplies?
Apparel, Consumer Books, CDs, Other
Footwear Electronics Jewelry, Toys
Product and PCs Theme Stores
Categories and Multiple Product
Categories Davids
Bridal CompUSA Borders Group
Ahold Footstar Office Depot Bulgari
Christmas Tree Shops Gap Inc. Radio Shack
The Disney Store CVS GH Bass Staples
Tiffany Co. Federated Maurices the
good guys! TransWorld Enter HE
Butt Nine West Tweeter etc Warner Bros.
Iceland Frozen Foods The Limited Zany Brainy
JC Penney World Marks
Spencer Zara QVC
Sears
Marshall Fisher, Ananth Raman Anna McClelland,
Rocket-Science Retailing is Almost Here Are
You Ready? Harvard Business Review July August
2000
3
Right Product in Right Place at Right Time and
Right Quantity
Inventory
Accurate Forecasts
Responsive Supply Chain
  • Accurate, accessible data on sales and inventory
  • Right incentives for buyers, suppliers and store
    personnel

4
State of retail supply chain management today
31
Too many of the wrong products
Department store markdowns as a percentage of
sales
26
21
16
11
6
1970 1980 1990 1995
5
Demand Forecasting
  • Blending Human and Artificial Intelligence

6
Obermeyer 2 product color slide

7
Early Sales is an Accurate Predictor of Lifecycle
Demand
Expert Forecast by a Committee of Four
Merchandisers
Forecast Obtained by Extrapolating the First 2
Weeks (11) of Orders
Average Forecast Error is 55
Average Forecast Error is 8
8

Season Sales vs Initial Forecast
Season Sales vs Forecast after first
32 of demand
2500

2500
2000
2000
1500
1500
Season Sales
Season Sales
1000
1000
500
500
0
0
0
500
1000
1500
2000
2500
0
500
1000
1500
Forecast after First 32 of Demand

Initial Forecast
9
Raw POS History
10
Raw Sales from January-July 01
The Thud!
11
Next, deseasonalize the data
12
Finally, correct for of stores
13
The Thud is a Gentle Slope!
14
Forecasting Ad Sales
  • Sales to a supermarket chain for a distributor

15
Forecast Errors by Sales Category
  • Mean Absolute Percentage Error

16
Expert Forecast for sample C Items
Booked represents human forecast. Billed
represents sales.
17
Implications for Forecasting
  • Human Machine wins over either operating alone.
    Evidence in First and Second Blended Forecasts.
  • Leave A items to human experts. The various
    algorithms tried so far all perform poorly
    (evidence of tacit knowledge?).
  • Automate forecasts for B and C items. Make it
    harder for humans to alter these forecasts.

18
Forecasting Forecast Error
  • Dispersion Among Experts

19
Using expert forecasts to gauge confidence in the
forecast
20
When experts agree, theyre more accurate
High Error
Average Error 116 units
Average Error 645 units
Low Error
Low Agreement
High Agreement
21
Inventory Planning (Hedging) Strategies
  • Benchmarking your performance, planning for short
    and long lifecycle products

22
Efficient Frontier
  • Shape and location of efficient frontier for the
    entire chain depends on
  • Number of doors
  • Variability of demand
  • Draw policy

23
Improving Planning for Replenishment Products
  • The Math Matters!
  • Look for low-hanging fruit in your sales data.

24
On Hand inventory prior to 4R-IPMax
4R
Ctrl
25
After implementing 4R-IPMax weeks of supply
relative to peer SKUs decreased 6.9
4R
Ctrl
Start of pilot
26
Lost sales prior to 4R-IPMax
4R
Ctrl
27
Lost sales for 4R-IPMax controlled SKUs decreased
from 6.6 to 2.2
4R
Ctrl
Start of pilot
28
Improving Planning for Replenishment Products
  • Highlight problems with stockouts and lost sales.
  • Beware of bullwhip effect.
  • Additional Profit Potential could be 1-5 of
    sales. Typically, a very quick hit and results
    are usually provable.

29
Supply Chain Planning for Short Lifecycle (and
new) Products
  • Understanding and exploiting demand uncertainty

30
Reading and reacting to early sales
100.0
This portion of demand can be supplied based on
an accurate forecast.
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Lead Time
Week of season
Replenishment Supply
Initial Shipment to Cover Read/ React Period
Safety Stock Based on Error Margin in the
Forecast
Read Market Order More of Hot Sellers
31
Improving Supply Chain Planning for Short
Lifecycle Products
  • Forecast (as realistically as possible) and share
    forecasts as broadly as possible.
  • Provide estimates of demand uncertainty as well
    use in scheduling.
  • Provide forecast updates as sales data become
    available.
  • Additional Profit Impact 2-20 of sales.
Write a Comment
User Comments (0)
About PowerShow.com