Understanding Leveraging - PDF document

1 Understanding & Leveraging Demand Varia
Understanding & Leveraging Demand Variability Rick Blair Who am I? Rick Blair VP, IBP Solution Consulting One Network Enterprises • 11 years in supply chain consulting • 18 years in ope

2 rations management Experience • Proces
rations management Experience • Process & Solution Design • IBP, S&OP, Demand Planning, Supply Planning • Statistical Forecasting • Operations & Materials Management • Procurement 2

3 Let’s start with some numbers In Major
Let’s start with some numbers In Major League Baseball… • W hat is a typical batting average? (number of hits divided by at bats) • What is a typical fielding %? (number of times

4 a defensive player properly handles a b
a defensive player properly handles a batted or thrown ball vs total opportunities) 3 25.4% 98.5% Variability Unpredictability 4 Which Product Would Y ou P refer to Manage? Demand Variabi

5 lity: an opportunity to differentiate yo
lity: an opportunity to differentiate yourself from competitors 6 Causes of Demand Variance Consumer behavior Seasonality New Product Introductions Promotions P rice changes Macro - economic

6 factors Insufficient supply (can’t bu
factors Insufficient supply (can’t buy if no stock on shelf) B2B customer behavior What Should We Do? I nfluence demand or adjust to it? …maybe both Demand Shaping Demand shaping is the

7 practice of influencing demand to bette
practice of influencing demand to better align with available supply. Typical levers include: • Price reductions (or increases) • Advertising • P roduct substitution • P roduct plac

8 ement, such as end of aisle or near che
ement, such as end of aisle or near checkout station Demand Agility Techniques for demand - driven flexibility and responsiveness: • Demand sensing • Fast replenishment, shorter lead tim

9 es • Supply chain network collaboratio
es • Supply chain network collaboration • Measure and improve forecast accuracy Demand Sensing Goal: isolate or predict demand shifts earlier than normal sensors recognize and adjust â€

10 ¢ Marketing/Big Data: may monitor social
¢ Marketing/Big Data: may monitor social media to pick up on trends that could influence consumer buying • Point - of - Sale (POS) demand shifts trigger quick re - planning of requirements

11 Automated Demand Sensing & Replenishmen
Automated Demand Sensing & Replenishment Supply Chain Network Collaboration 14 Multi Party Networks E nable Real - Time C ollaboration Legacy Legacy Legacy Legacy Legacy Legacy Legacy Legacy

12 ERP / EDI Single Multi - Tenant Insta
ERP / EDI Single Multi - Tenant Instance Cust 1 Cust 2 Cust 3 Cust 3 Cust 3 Hub Hub Hub Hub Hub NETWORK PLATFORM Single Multi - Tenant Multi Party Instance Cust1 Cust 2 Cust 3 Cust 4 Cus

13 t 5 Legacy Legacy Legacy Legacy Legacy L
t 5 Legacy Legacy Legacy Legacy Legacy Legacy Legacy Legacy Service Center Legacy Supplier Supplier Legacy Partner Service Center 3PL / Carrier Prime Vendors Allies Legacy Legacy Legacy 3PL

14 / Carrier 3 rd Party Service Suppliers
/ Carrier 3 rd Party Service Suppliers Supply Chain Visibility Upstream & Downstream Minimizes or Eliminates Bullwhip Effect 15 Why Measure F orecast A ccuracy? What is Forecast Accuracy?

15 • A measure of deviation between plan
• A measure of deviation between plan and actual • Less deviation = greater accuracy What motivates an organization to track accuracy? • Hold individuals and groups accountable • Be

16 nchmark against other companies • Mana
nchmark against other companies • Manage inventory levels • Improve management of business How is MAPE Calculated? Mean Absolute Percent Error = MAPE = ∑│PE│/N • N = number of peri

17 ods for which we have PE values • │P
ods for which we have PE values • │PE│ = absolute value of the PE (Percent Error) Weighted MAPE = average of individual MAPEs weighted by actual shipments = ∑item MAPE * │(item actu

18 al / total actual)│ Let’s consider
al / total actual)│ Let’s consider an example… MAPE Example 1. Error: The difference between Forecast and Actual 2. Absolute Error: Convert negatives to positives 3. Absolute Percent

19 Error = Absolute Error / Actual MAPE Exa
Error = Absolute Error / Actual MAPE Example Mean Absolute Percent Error (MAPE) – Average of APEs for multiple items or multiple periods or both MAPE = 30.5% = A verage of 10 SKU APE value

20 s MAPE for Family A: At Family level = 2
s MAPE for Family A: At Family level = 2.1% At SKU level = 30.5% WMAPE Example Weighted Mean Absolute Percent Error (WMAPE or WAPE) – Average of individual MAPEs weighted by actual shipments

21 WMAPE = 5.0% = Weighted Average of 10
WMAPE = 5.0% = Weighted Average of 10 SKU APE values WMAPE for Family A: At Family level = 2.1% At SKU level = 5.0% MAPE & WMAPE: Impact of Small Changes What if SKU5 Actual was 2 (not 1)? M

22 APE changes from 30.5% to 15.5% WMAPE ch
APE changes from 30.5% to 15.5% WMAPE changes from 5.0% to 4.9% A one unit change can have a huge impact on MAPE. WMAPE is less sensitive. MAPE & WMAPE: Zero Actual Values What if SKU5 Actu

23 al was 0 (not 1)? MAPE changes from 30.5
al was 0 (not 1)? MAPE changes from 30.5% to 10.5% WMAPE changes from 5.0% to 4.9% Pay close attention to instances where Actual = 0. D ividing by zero causes an error, so the MAPE formula

24 may assign 0%. This can be misleading
may assign 0%. This can be misleading since 0% should represent zero deviation between forecast and actual. Recommendation: Analyze Top 10 WMAPE Items In Summary: Are You Asking These Ques

25 tions?  Does demand variance give you
tions?  Does demand variance give you a competitive advantage?  What’s your strategy: Shape demand or be agile?  Can you shorten lead times and replenish faster?  Is your S&OP st

26 uck inside your four walls? Do you colla
uck inside your four walls? Do you collaborate effectively throughout your supply chain, in real - time?  Are you measuring forecast accuracy to pin blame or improve the company? Thank Yo