Intelligent Supply Chain

Similar documents
Data Science & Big Data Practice

Delivering new insights and value to consumer products companies through big data

MSD Supply Chain Programme Strategy Workshop

Steel supply chain transformation challenges Key learnings

Maximising supply chain throughput with existing infrastructure

Supply & Demand Management

Model, Analyze and Optimize the Supply Chain

Aspen Collaborative Demand Manager

Improve the Agility of Demand-Driven Supply Networks

White Paper February IBM Cognos Supply Chain Analytics

How to Cheat and Make Better Decisions with Predictive Analytics. Track 1 Session 3

Tapping the benefits of business analytics and optimization

Supply chain intelligence: benefits, techniques and future trends

Four distribution strategies for extending ERP to boost business performance

4 Key Tools for Managing Shortened Customer Lead Times & Demand Volatility

Data Science & Big Data Practice

Supply Chain development - a cornerstone for business success

Vehicle Sales Management

Supply Chain Acceleration: Our Offering for Enabling Growth

How IT Can Help Companies Make Better, Faster Decisions

Supply Chain Management Build Connections

Overview, Goals, & Introductions

BI STRATEGY FRAMEWORK

IoT Changes Logistics for the OEM Spare Parts Supply Chain

EVERYTHING THAT MATTERS IN ADVANCED ANALYTICS

Making Machines More Connected and Intelligent

Staying a Step Ahead by Comprehensive Industrial MRO Outsourcing

Optimizing Inventory in Today s Challenging Environment Maximo Monday August 11, 2008

Oliver Wight Sales & Operations Planning and Demand Management as a Competitive Differentiator

The Supply Chain Excellence Study Long version

Strategies for optimizing your inventory management

Business Intelligence Meets Business Process Management. Powerful technologies can work in tandem to drive successful operations

Leverage the Internet of Things to Transform Maintenance and Service Operations

Predictive and Prescriptive Analytics An Example: Advanced Sales & Operations Planning

Business Challenges. Customer retention and new customer acquisition (customer relationship management)

Introduction to Strategic Supply Chain Network Design Perspectives and Methodologies to Tackle the Most Challenging Supply Chain Network Dilemmas

8% of US GDP (USD 1.2 Tn) is Services. $ from Service Sale = 4*$ from product sale. Business drivers for shift in focus towards Aftermarket

Accenture NewsPage Sales Force Automation: Empower your people

Data Science & Big Data Practice

China Grand Auto: Partnering with SAP on a State-of-the-Art Platform for a Multibrand Dealer Group

Supply Chain Performance: The Supplier s Role

Industrial Roadmap for Connected Machines. Sal Spada Research Director ARC Advisory Group

Big Data for Investment Research Management

Big Data Strategies Creating Customer Value In Utilities

ORACLE SUPPLY CHAIN AND ORDER MANAGEMENT ANALYTICS

pg. pg. pg. pg. pg. pg. Rationalizing Supplier Increases What is Predictive Analytics? Reducing Business Risk

Amplify Serviceability and Productivity by integrating machine /sensor data with Data Science

Maintenance, Repair, and Operations (MRO) in Asset Intensive Industries. February 2013 Nuris Ismail, Reid Paquin

Accenture NewsPage Distributor Management System: The engine behind your business

Deliver a Better Digital Customer Experience Through Sonata s Digital Engagement Solutions

What s Trending in Analytics for the Consumer Packaged Goods Industry?

Endeavour Dynamics Offering

Fleet Optimization with IBM Maximo for Transportation

msd medical stores department Operations and Sales Planning (O&SP) Process Document

How Organisations Are Using Data Mining Techniques To Gain a Competitive Advantage John Spooner SAS UK

Logistics Management SC Performance, SC Drivers and Metrics. Özgür Kabak, Ph.D.

Integrated Sales and Operations Business Planning for Chemicals

Managing End-to-End Supply Chain Costs in a Down Economy

Big Data - An Automotive Outlook

SALES AND OPERATIONS PLANNING BLUEPRINT BUSINESS VALUE GUIDE

White Paper May Seven reports every supply chain executive needs Supply Chain Performance Management with IBM

Sonata Managed Application Lifecycle Services

QlikView for Supply Chain. Chemical and Mill Products

Maximizing Returns through Advanced Analytics in Transportation

decisions that are better-informed leading to long-term competitive advantage Business Intelligence solutions

Enhancing Sales and Operations Planning with Forecasting Analytics and Business Intelligence WHITE PAPER

SAP BUSINESSOBJECTS SUPPLY CHAIN PERFORMANCE MANAGEMENT IMPROVING SUPPLY CHAIN EFFECTIVENESS

Grow Your Business, Serve Customers Better and Save Costs with Logistics Analytics

RapidResponse Inventory Management Application

Driving Insurance World through Science Murli D. Buluswar Chief Science Officer

Streaming Analytics and the Internet of Things: Transportation and Logistics

Transportation Management

Load Building and Route Scheduling

Strategic Data Governance

abf Avercast Business Forecasting The Trusted Name in Demand Management. Software Features: Enterprise Level Software Solutions for: The Cloud

SAP Solution Brief SAP HANA. Transform Your Future with Better Business Insight Using Predictive Analytics

Inventory Cycle Counting

SOLUTION OVERVIEW SAS MERCHANDISE INTELLIGENCE. Make the right decisions through every stage of the merchandise life cycle

Building a Business Case for Supply Chain Execution in the Cloud

Balancing Uptime and Working Capital: Maintenance and Inventory Strategies in Mining

Supply Chain Simulation: Why Its Time Has Come

Use Advanced Analytics to Guide Your Business to Financial Success

Plan forecast optimise

MWPVL. Leadership in Supply Chain and Logistics Consulting. Options to Improve Productivity at a Parts Distribution Center

BUILDING A DEMAND-DRIVEN SUPPLY CHAIN THROUGH INTEGRATED BUSINESS MANAGEMENT FROM EXECUTION TO OPTIMIZATION ADVANCED EXECUTIVE EDUCATION SERIES

Supply Chain: improving performance in pricing, planning, and sourcing

QlikView for Supply Chain. High Tech

T r a n s f o r m i ng Manufacturing w ith the I n t e r n e t o f Things

A PRACTITIONER S VIEWPOINT

A Forrester Consulting Thought Leadership Paper Commissioned By Zebra Technologies. November 2014

Baker s Dozen: 13 Ways Process Intelligence Drives Supply Chain Value

Value Creation Through Supply Chain Network Optimization To Address Dynamic Supply Chain

MES and Industrial Internet

Transcription:

Data Science & Big Data Practice INSIGHTS ANALYTICS INNOVATIONS Intelligent Supply Chain Enable Better Decision Making Through Effective use of Data

Supply Chain is generating significantly large unstructured data SCM Big Data Analytics is the process of applying advanced analytics techniques in combination with SCM theory to datasets whose volume, velocity or variety require information technology tools from the Big Data technology stack; leveraging supply chain professionals with the ability to continually sense and respond to SCM relevant problems by providing accurate and timely business insights. SCM Data Volume and Velocity vs. Variety Challenges in Supply Chain Components, equipment and finished goods travel through multiple channels, and increasingly, across multiple time zones. A foul-up by a truck driver in Zimbabwe can cost a corporation in New York millions of dollars if the situation isn t handled correctly The scale, scope and depth of data, supply chains are generating today, is accelerating, providing ample data sets to drive contextual intelligence It is clear that the majority of supply chain data is generated outside an enterprise Forward-thinking manufacturers are looking at big data as a catalyst for greater collaboration Source: Conference Paper, Big Data Analytics in Supply Chain Management

..creating huge demand for Big Data Analytics among leaders Disruptive Technologies for Supply Chain Supply chains evolve into value webs Big Data Analytics Linear supply chain are evolving into. complex, dynamic and connected value webs Digital Supply Chain COMPETITION COMPETITION Internet of Things Cloud Computing Advanced Robotics 3D Printing Drone/Self-guided Vehicles Sharing Economy (e.g. Uber, Airbnd, Instacart) Disruptive & Important Interesting, but unclear usefulness Irrelevant % of respondents N = 1057 64% of supply chain executives consider big data analytics a disruptive and important technology, setting the foundation for long-term change management in their organizations. Value is based on the production of goods and services Value is based on knowledge exchange that drives proactive production of goods and services Analytics enables more complex supplier networks those focus on knowledge sharing and collaboration as the value-add over just completing transactions Sources: Deloitte Research & SCM World

Supply Chain Challenges you face How to manage supply chain volatility and demand uncertainty? Product Development Procurement Operations Wear housing & Distribution How to Optimize product features Optimize production level Concept testing De-risk supply chain Improve accuracy of raw material price prediction Rate vendors objectively Reduce supply variability Schedule production based on demand Increase production with available resources Reduce machine downtime Identify failure proactively Improve channel efficiency Reduce total distribution cost Optimize route for material distribution Reduce cost of distribution How to integrate sales and operations planning?

DSI Supply Chain Solution Framework Our Supply Chain Solution framework is developed around complex mathematical models and IP ed algorithms These models take into account a wide range of variables, such as the additional costs due to variations in the speed with which different suppliers can deliver their goods, proximity of the delivery centres Information about consumer needs and wants to develop new product and/or brand extension internal/external data to build pricing models that maximize profit margins Demand of product by location, by customer type, product availability by warehouse, etc. Users can select options with the highest return and the lowest investment to maximize profit ANALYTICS Operations Management Machine Downtime reduction It helps users to make better purchase decisions, more flexible capacity planning, accurate demand forecast and identify optimum route of distribution. Our Supply Chain Big Data framework is highly scalable and our Data Scientists are currently working on sourcing data from connected technologies and Internet-of-Things to optimize the supply chain of the future

How do you gain How Data Science helps Supply Chain Leaders Improvement in customer service & demand fulfilment of 10% or greater Faster and more effective reaction time to supply chain issues Increase in supply chain efficiency of 10% or greater Greater integration across the supply chain Optimization of inventory and asset productivity More effective S&OP process and decision making Improved cost to serve Better customer and supplier relationships Improvement in customer service and demand fulfilment of less than 10% Increase in supply chain efficiency of less than 10% Improvement in demand driven operations Shortened order-to-delivery cycle times Embedding big data analytics in operations leads to a 4.25x improvement in order-to-cycle delivery times, and a 2.6x improvement in supply chain efficiency of 10% or greater How Supply Chain Leaders gaining by deploying Data Scientist Team of data scientist Traditional database personnel Shortened order-to-delivery cycle times Improvement in demand driven operations Increase in supply chain efficiency of 10% or greater Better customer and supplier relationships Improve cost to serve More effective S&OP process and decision making Faster and more effective relation time to supply chain issues Greater integration across the supply chain Improvement in customer services and demand fulfilment of 10% or greater Optimization of inventory and asset productivity Companies that employ a team of data scientists are far more likely to generate a range of important supply chain benefits from their use of big data analytics Source: Accenture Research

Transforming your Data Chromosome DSI Case Studies on Supply Chain Analytics 1 10 01 1

80.0% 85.0% 90.0% 92.0% 94.0% 95.0% 96.0% 97.0% 98.0% 98.5% 99.0% 99.5% 99.7% Inventory Profit Margin Spare Parts Inventory Optimization The client is a global leader in the design, manufacture, installation, and maintenance of wind, driven power generation plants.. It had been consistently carrying a large inventory to ensure high service level as agreed with its customers. The client wanted to reduce its inventory without affecting the service level. It had been using an automated forecasting and inventory management tool to keep track of inventory. Order used to get generated automatically whenever existing stock fell below the required stock based on pre defined inventory parameters like safety stock, lot order quantity, etc. The objective of the project was to build up an inventory management model based on maximization of RoCE. Solutions Focus was on high value, capital spares both repairable and non-repairable Determined impact of lead time reduction in overall inventory reduction and reliability improvement and identified the critical items, where small improvement could lead to large gain Identified environment parameters affecting parts failure Inventory segmentation and optimization of service level for each different class of inventory What-if scenario analysis between service level and RoCE Business Questions Which are the high cost critical spares contributing to large inventory cost. What is causing high inventory for these spares Which critical spares are more likely to fail What environmental conditions (wind speed, humidity, temperature) are more likely to cause frequent failure of parts How much additional spares do I need to increase service level from x% to y% Trade off between predicted spares stock (RoCE) and likely achievable service level Value Proposition Optimize stock levels, reduce spares costs and increase service level; Impact Reduction in inventory by approximately 20% with increased reliability Automatic alert before a component failure reduced the need for ongoing inspections Managed spares requirement based on weather forecast by location Better forecast of parts failure helped in pro-active spares management and higher reliability Reliability Type-A Type-B Type-C 0.19 0.17 0.15 0.13 0.11 0.09 0.07 0.05-7 DII -3 DII -10 DII Current State 0.5 0.8 Asset 1 Turnover 1.5 2 2.5 3 3.5 10% RoCE 15% RoCE 20% RoCE

Route and Fleet Optimization The client is a third party logistics service provider supporting transportation requirement of customers from varied industries like mining, manufacturing, retail and courier services. The client provided transport solution from its 300+ branches across one of the largest geographies in Asia. 25-20% of transportation requirement is serviced by its own fleet while the rest is serviced using hired vehicles from the market. It wants to utilize its own fleet judiciously into those routes where return is maximum and where availability of outside vehicle is lower. Another business constraint for the client was preference of drivers to ply on the same route because local knowledge was critical to manage operations efficiently. The objective of the project is given below: Demand projection by route by product type (steel, cement, furniture, oil, etc.) by major customers for next 6 months Identify optimum combination of own fleet and third party vehicles by route for the next 6 months. The combination of fleet includes type of vehicles (6 wheelers, 10 wheelers, oil tankers, etc.). Provide optimum route by type of vehicle for long term contracts Business Questions What is the optimum combination of own fleet and market hired fleet by route? What combination of fleet types will manage demand variability with lowest cost? How to trade off between penalty for late delivery with cost of transportation? How to trade off between smaller capacity trucks to gain flexibility and higher capacity truck to gain cost benefit Value Proposition Optimum combination of types of vehicles, number of vehicles to manage demand variability with higher service level at lower cost Solutions The scope of the project was made limited to those top routes which were providing 80% of the revenue to the client. Pareto rule was again applied to select only those top products which were generating 80% of the revenue on the selected routes. The type of services were segregated into a) contract type and b) on-spot depending upon the type of payment and frequency of repeat business. It was also discovered that the contract type business needed the deliveries to be made at the same locations in majority of the cases. Following Visualizing demand: A baseline model was developed to represent the existing route structure and flow. Each route was scored based on demand volatility over the months Shipment consolidation: Identified the opportunity in terms of consolidating multiple demand generating at the same time either by using bigger vehicles or by consolidating nearby routes. Freight spend: Integer programming and other meta heuristic algorithms were used to optimize LTL (less-than-truckload) and TL (truckload) spend. Scenarios were used to create a set of sensitivity analyses of TL/LTL utilization thresholds. Impact Distribution spend was reduced by 15 percent 12% improvement in ontime delivery through load balancing Identified opportunities to increase truckload shipments and reduce LTL spend.

Demand Forecasting The client is one of the largest engine manufacturing and servicing companies. It has been managing multiple large Maintenance, Repair, and Overhaul (MRO) facilities across the world. The client engages with its customers both on long term contracts and on the spot contracts. On the spot contracts have been always more profitable than long term contracts. The client wanted to predict demand for engine servicing from on the spot contracts for a period of 3 years. This was useful for the client to sell long term contracts based on spare capacity by region. The demand for on the spot contract has been very volatile over the years. Predicting the same at a facility level has always been prone to large error. Time to time engineering upgrade of machine parts has also made prediction more difficult. The objective of the project was to understand the reason for demand variability by MRO facility. Business Questions How to trade off between assured long term contracts and more profitable on the spot contracts? How to distribute overall capacity utilization in an uniform manner throughout the year by reducing frequent spark in servicing demand? How to control inventory by reducing demand volatility? How to improve forecasting accuracy at a micro (product sub group) level? How to price my service for a long term contract based on future demand potential? Do I need to increase capacity in next 3-5 years? Improved forecasting accuracy helps in achieving higher margin and more uniform capacity utilization Solutions We deployed both time series analysis (Exponential smoothing, ARIMA, etc.) and causal based forecasting algorithms in the project. Macro economic factors like GDP growth, IIP growth etc. were considered to develop a demand forecasting model using both traditional multi parameter non linear regression techniques Machine algorithm techniques like Random Forest was also used along with bagging and boosting methodology Ensemble methodology was used to average out output from multiple forecasting models Expert s comments sourced from different industry websites, blogs, annual reports, etc. were also analysed using Natural Language Processing (NLP) techniques to improve model accuracy Multiple What-if scenario analysis were created based on probabilistic demand forecasting model Impact Forecasting accuracy improved from below 60% to 75%+ in the first 6 months of implementation Capacity utilization increased by 10% across all MRO facilities Safety stock reduced by more than 40% due to higher accuracy of forecast Overtime cost reduced significantly across many MRO facilities. Value Proposition 30% 0 80% 70% 60% 50% 40% Mnth-0Mnth-1Mnth-2Mnth-3Mnth-4Mnth-5Mnth-6 Fcst Accuracy (LHS) 4 3.5 3 2.5 2 1.5 1 Safety Stock (RHS) 0.5

100.00 93.35 98.94 92.90 110.77 Machine Downtime Analysis A leading DI pipe manufacturer in India produces pipes of varying diameters ranging from 80-800mm, each of two different types (K7 and K9). The company management experienced a considerable amount of gap between the ideal number of pipes that could be produced and the ones actually were produced and delivered. The gap or the loss was attributed to three factors Downtime Rejection Performance Solutions For each stretch of time that a machine spends in a state, all pertinent information was collated from different data sets and brought into a unified representation Events were recorded as Scheduled and Unscheduled, and further divided as Mechanical, Electrical and Operational events. Analysis was separately undertaken for each type of event and each machine. Metrics like frequency, mean time between failures, mean time to repair, frequency distribution of time to repair, etc were computed for all top events Business Questions Which are the critical machine / components failing more frequently How much idle time is lost due to failure of the machines / components How can we predict failure based on operational parameters How to find relationship between failure of one component with failure of other components How much time is spent on repair of failed machines / components How can we reduce overall downtime of machines / components How machine / component failure is affecting product quality / rejection rate Value Proposition The mandate of this project is to bring down overall loss by 15% and reduce rejection rate by more than 50% Impact Overall machine downtime reduced by little over 16% Finished goods production increased by 11% Rejection rate came down by more than 50% Dealer Segmentation 115.00 Q4 FY14 indexed to 100 110.00 105.00 100.00 95.00 90.00 85.00 80.00 MTBF (hrs) for CCM Mould Change Scheduled

Value Index Vendor Rationalization One of the leading electronics manufacturing companies having manufacturing and assembling facilities spread across the globe had distributed sourcing functions by product category. The sourcing team of each product category had overlapping vendors, SKUs, multiple contracts with the same vendor with numerous different pricing structures and payment terms. The client realized financial (higher inventory, non-optimized delivered cost, etc.) and non-financial (complex reconciliation process, etc.) impacts of the overwhelming complex vendor portfolio. In an endeavour to simplify the sourcing process and to reduce overall cost, the client embarked on the journey to centralize the material procurement department across product categories. Europe region was selected to kick start the centralization process because of higher per unit cost in that region. Solutions SKU selection: Classical inventory analysis like ABC analysis and VED analysis was done to classify SKUs based on numerous parameters like profit, revenue, quantity, etc. SKU commonalization: All vendors contributing 90% of the overall supply of the selected SKUs are identified and a consolidated master list was created Vendor scorecard: All vendors in the master list were evaluated based on numerous KPIs like cost, quality, service level, etc. An weightage system was developed through monetisation (dollar value) of each parameter. Example: 1 day delay in delivery reduces profit of the company by xx USD Vendor optimization: Optimum number of vendors for each SKU was determined based on volume, value, criticality, etc. of the SKU in the overall business Business Questions What is the right number of vendor for important SKUs? How to trade off risk of overdependence on few vendors with too many vendors proving less than optimum cost efficiency? Which SKUs can be grouped together based on product engineering? Value Proposition Optimum number of vendors that reduce overall cost of procurement, improve material availability, reduce risk and improve quality of raw material Impact Aggregate raw material cost delivered at all facilities in Europe reduced by 7% during the first six months of implementation. The annual potential cost saving was determined at 12% after stabilization. Average raw material inventory reduced by 8 days of sale. This was primarily achieved through commonalization of SKUs across product categories 100 80 60 40 20 0 Critical Vendors 0 20 40 60 80 100 Vendor Scorecard

Transforming your Data Chromosome DSI Distribution Channel Efficiency Enhancement Tool

Product DSI Channel Efficiency (DSI-CE) Tool DSI channel member segmentation tool evaluates a dealer s performance across multiple dimensions (Volume, Credit, Stickiness, Differentiation, Premium). The Big data based machine learning algorithm identifies the key parameters driving dealer performance. Some key Questions 1. How do I rate my dealer s performance? DSI-CE Architecture Insights 2. How do I utilize maximum information from existing data sources for my decision process? 3. How do I get actionable insight about each segment of best performing and worst performing dealers? 4. How my customers and dealers are perceiving my competitors? Benefits Help you to decide the optimum business parameters to set right targets for channel members Provide insight on the sentiments of your customers and dealers Streamline the complete Distribution value chain Dealer / ECA Scoring Tool (R-CAP) Sales, Stocks Data Channel Management Tool (CMT) Open Source, Customized Statistical Modelling & Machine Learning Common, Unified, Dynamic Big Data Repository Planning Data Smart Incentives Design Tool (SIDT) Market Data Sales Forecasting Tool (SFT) Other Structured Data Sources Payment Cycle Analysis Tool (PCAT) Excel Reports Customer Feedback Analysis Tool (CFT) Open Source/ Unstructured Data The DSI-CE business suite has a customized interactive Big Data based Decision Support System (BDSS) that help to rate, track, compare performance of distributors, dealers/retailers and B2B customers This is supported by a detailed solution architectures, big-data integration framework, statistical and machine learning algorithms, data visualizations and decision support tools

Product DSI Channel Efficiency (DSI-CE) Tool Advanced statistical and machine learning algorithms are used to establish relationship between dependent variables (dealer growth) and hundreds of independent variables (shop details, promotional activities, etc.) Identification of Important Parameters Overall Tenure 0.35 Associated prdts COV of sales # of BTL activities Project / Ops On-time delivery 0.45 Dealer Rating Credit period Store size Signage 0.20.. New biz ratio Neural Network for identifying important variables affecting performance Dealer1 Dealer2 Hierarchical clustering for segmentation of salesperson

Product DSI Channel Efficiency (DSI-CE) Tool CREDIT Dealer segmentation provides detail insight about each segment of dealers and their characteristics. This helps in taking the right action for each group of dealers under different business dynamics Segment Characteristics 1. Likely market reach is above 70% 2. Keep outstanding above 60 DOS 3. 32% of the stores are likely to grow above 15% in the next 12 months A B Dealer Segmentation VOLUME A B C D E 102 75 134 157 4. Sales is strongly correlated with outstanding provided to these dealers 5. SOB is below 25%. 6. Likely to switch allegiance to competitor brands 7. Adherence to company guidelines is below average C D E Nos represent number of dealers in this segment Characteristics about each micro segment (e.g., Volume High, Credit Low, Stickiness Medium) are populated to provide right business insight for future action

Product DSI Channel Efficiency (DSI-CE) Tool Volume Growth (%) Parameter Optimization 30 25 Identification of right parameters affecting dealer growth helps the organization to focus its action on key action points. In the next phase optimization of useful parameters is required to align business decisions with dynamic business environment 20 15 10 5 0 0 10 20 30 40 50 60 70 80 90 100 Share of Business (%) Optimization of the selected parameters by geography for selected segment of dealers is done to meet specific business goals during a specific business cycle

INSIGHTS ANALYTICS INNOVATIONS