Far too often, Supply Chain management leaders make decisions in a data vacuum. But that doesn’t work in today’s fast-paced market. Rob Van Driel (Solutions Consultant, Anaplan) explained why & how Supply Chain leaders need to make timely, value-based decisions so you can respond quickly to shifts in demand and customer needs. Because: when value is king, margins are optimized, and profit is maximized.
Presented by Rob Van Driel, Solutions Consultant Anaplan on Supply Chain 4.0 : ready to operate in the digital era? (29 Nov, 2018)
4. Driving a new age of
connected planning
2011
commercial launch
20 offices
in 13 countries
75%+ FY17
subscription
revenue growth
(FYE January)
1000+
employees
in 14 countries
900+ customers
in over 40 countries
Partners
Best-in-class
250+
apps
Patented
in-memory data
engine
$120M FY17
total revenue
momentum
6. Barriers that needs to be addressed:
Organizational
Silos
• Lack of visibility
• Sharing of Data
Data
• Inconsistency
• Quality
• Reconcile
Flexibility
• Dynamic changes
• Varying timeframes
• Different planning
levels
Technology
• Manually intensive
• Alignment tools
Workflow
• Different planning
horizons & cycle
times
• In ability to quickly
change pans
7. legacy vs Anaplan
ERP system
MES system
Transportation management system
Warehouse management system
MRP system
HCM tool
Microsoft (Excel®)
Financial planning system
CRM tool
S&OP
Management
reports
Inventory
management
MPS
Forecasting
tool
Strategic
pricing tool
Supply
optimization
tool
Order
fulfillment
tool
Single, secure
source of planning
and decision data
Greater
collaboration, deeper
insights,
faster alignment
Dynamic,
continuous planning
for any area of your
business
Across one
department or area
Across the company
One business process
9. “Anaplan gave us speed and agility.”
RK Del Rosario, Supply Chain Planning Manager
ANAPLAN FOR CONNECTED PLANNING
Days slashed from routine
planning processes
CHALLENGES
• Legacy tools took up to 6 hours to run some
operations
• Planners used inaccurate averages, resulting in
imprecise forecasts
• Process failures resulted in lost sales and excess
inventory
RESULTS
• Two-week planning process was cut to two days
• A five-day revision process now takes five
minutes
• Channel, SKU, and customer profitability are
available monthly
11. Step 1: connect supply chain planning data and processes
Transportation
management system
ERP system
MES system CRM tool
Warehouse
management system
Microsoft (Excel®)HCM toolMRP system
Financial planning
system
Benefits:
• Automated integration
• Single Repository
• Visual representation of data
• Network visibility
12. Step 2: Connect people and plans across the organization
Demand
Signal
ManagementTrade
Promotions
Management
Demand
Management
Planning
Dashboards
Collaborative
Planning
Statistical
Forecasting
Sales
Forecasting CRM
Pipeline
Management
NPI/EOL
Forecasting
Financial
Forecasting
Marketing
Forecasts
Demand
Shaping
Demand Analyst
Marketing
Account Management
Sales management
Demand Planning
Controller
Rough-Cut
Capacity
Plan
Supplier
Plan
Sourcing
Plan
Planning
Dashboards
Capacity
Plan
(Constraints)
Consensus
Demand
Plan Capacity
Plan
(Resources)
Allocations
Plan
Inventory
Plan
Materials
Planning
Master
Production
Schedule
Procurement
Plan
Production Manager
Supply Planner
Inventory Manager
Master Planner
Distribution Manager
Warehouse Manager
Consensus
Planning
Executive
S&OP
Demand Planning
Supply Planning
Decision
Benefits:
• Complete profitability analysis
• Better anticipate market changes
• Real time implications supply chain
decisions have on corporate strategy
• Improve service levels
13. Step 3: Collaboration Across The Network
Raw Material
Suppliers
Contracted
Producers
Transportation Distribution
Center
Warehouse Customer
Channel
CustomersProduction
Facilities
3rd Party
Logistics
Channel
Partners
Transportation
Benefits:
• Visibility across all extend network events
• Immediate action due to deviations extended network
• ‘What if’ scenarios including suppliers and customer
• Better support omni-channel
14. Connected Planning Creates Value
Plan demand for tens of thousands of
SKUs globally
Consumer Products Manufacturer
Reduced Planning Cycle Time 80% (2
weeks to 2 days)
Food & Beverage Manufacturer
Improved forecast accuracy by 15%
Food & Beverage Manufacturer
Reduced functional FTE requirement for
impacted processes by 40%
High Tech Manufacturer
Reduced Order Processing Time 70%
from 7 to 2 Days and increased capacity
by 150%
High Tech Manufacturer
Effectively Planned 55% incremental
demand and revenue via new bundles
High Tech Manufacturer
1.5% sustained increase in Net
Income
Global Apparel Manufacturer
Reduced inventory on hand by an
estimated $100M
Global Apparel Manufacturer
Increased EBIT 25+% over 3 years
Global C&IP Manufacturer
Intelligent
Self-learning, Insightful,
Predictive, Cognitive
Collaborative
Networked, Inclusive,
Distributed, Accessible
Dynamic
Real-time, Responsive,
Flexible, Fast
16. Intelligent Planning Roadmap
Optimization
• Fully integrated into
Anaplan UI
• Optimal – Feasible problem
• Gurobi Solver Engine
Machine Learning/AI
• POCs with Google Cloud and
other partners
Predictive
• Currently available 26
algorithms
• Examples: Linear regression,
Exponential smoothing,
Erlang, Holt-Winters
Optimization
• Future versions to include all
major problem types
• Multivariable linear
regression
Machine Learning/AI
• Leverage ML/AI algorithms
and models for planning data
Immediate Term Future
This content is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decision.
17. Connected
Network
Planning
• Collaborative Planning across the supply network
• Dynamic and Continuous Planning and Optimization
• Intelligent, Faster and better decisions
Machine
Learning
Algorithms
Continuous
Optimization
Leading Edge Innovative Technology
18. Component of Function
that can be changed.
e.g., unit volume
Optimization Problem Definition
Objective Variable Constraint
Maximize or minimize
the value of some
function, F(x1,x2…xn)or
Determine Feasibility
e.g., maximize profit
Constraints on
individual x’s and/or
combinations of x’s
e.g., production
capacity
19. Using only model based line
items and formulas, the
optimizer does not require any
specific modeling skills.
Optimizer Highlights
UI DRIVEN MODEL BASED FAST CALCULATION
Problem definition relies on a
simple, Anaplan standard UI to
give more flexibility to users.
Gurobi is the fastest
optimization engine on the
market. Because the operation
locks the model, it’s a
prerequisite for the optimizer
to be fast.
This content is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decision.
20.
21. ML POC: Large Beverage Company
POC Hypothesis: Determine if forecast improvements could be achieved
to yield COGS savings and associated SCM benefits
Scope: Apply ML on POS data over 8-10 weeks
Result: ML yielded a 15% improvement in forecast accuracy projected
to save $2M in COGS for the sample of products in the study.
Benefits: Targeted areas to apply the result
• Financial forecasting at granular levels
• Better allocate marketing spend across brands
• Trade promotions and marketing event planning that lift sales up
Additional downstream benefits through SCM
• Reduced Safety Stock Investment
• Reduced Expired and Obsolete Products
• Reduced Working Capital
22. ML POC: large CPG/consumer healthcare company
Result:
• more accurate forecasts than stat models in 79% of cases
• On average ML yielded 24% MAPE (vs. industry avg. of 36-40%)
• potentials of $4.1M in savings.
• $1.7M by addressing some of the missed sales
• $2.4M in efficient inventory management
Opportunities: Operating cost savings with improvements in sales prediction
by:
• Forecasting at granular level
• Optimizing supply planning on top of ML results
$4.1MSavings for 6 weeks in forecast accuracy
24. ML Model Performance
(Illustrative with one product example)
• ML models (with any combination of data) performing consistently better than stat models
• Granularity of forecasting ability (at UPC-DC level) shows models can be tuned to yield better result for
individual product or brand
25. Anaplan PlatformUsers Google Cloud Platform Data Sources
POS
ML POC Solution Architecture
Financial
Forecasting
Demand
Planning
Others
Data
Hub
External
Data
Shipping
Promotions
Big Query
Machine
Learning
Cloud
Storage
26. 26
• A connected planning approach results in better, more
collaboratively created plans that are resilient in the
face of change
• Anaplan is a unique platform to develop, integrate,
reconcile, refine, and manage plans
• Anaplan capabilities can be implemented quickly using
agile processes
• Anaplan is highly scalable and supports very complex,
financial, product and supply chain planning models
Key takeaways
Questions