Building & Leveraging the Metrics Framework for driving Enterprise Performance Management (EPM) A Supply Chain Management view

Transcription

1 Building & Leveraging the Metrics Framework for driving Enterprise Performance Management (EPM) A Supply Chain Management view Tejas Faldu, Srikanth Krishna Abstract Failure to design a relevant and integrated metric framework greatly reduces the capability of business intelligence systems to provide the right insights for effective decision making. This paper outlines how a metrics framework based approach to enterprise performance management is a fundamental step to leverage investments in business intelligence and analytics. 11 December 2005

2 According to an AMR Research report, EPM spending in 2005 was estimated to be close to $22.2 Billion of which more than $9.5 Billion is being spent on Business Intelligence systems and Analytic applications. Armed with such sizeable budgets, enterprises are actively planning strategies to infuse analytics everywhere in their business. As enterprises undertake this analytics journey, they would need to answer some important questions. Some of them are - Do all roles understand the financial impact of their performance on the overall enterprise performance? Is there an enterprise wide awareness and understanding of the overall strategic and financial objectives of the enterprise? Are all processes and roles of your enterprise mapped to some key metrics which determine the success of their performance? Can information be easily accessed and analyzed quickly to take right decisions at the required time? Is there a mechanism to periodically review actual performance measures and also redefine performance measures in the changing business context? Is there an integrated single view of performance across functions and across hierarchies? The answers to these questions reside in establishing an enterprise performance metrics framework. Unless such a framework is built, investments in analytics and business intelligence systems will give sub-optimal returns. In this paper we demonstrate a process-oriented approach for establishing and using an enterprise performance metrics framework using the illustration of how this is achieved for Supply Chain Management in the CPG industry. Business Case for Supply Chain Performance Management Supply Chain Performance Management has been a critical focus area for leading CPG players in their pursuit of developing agile, lean and efficient customer oriented supply chains. One of the biggest challenges that CPG players face today is to maintain the delicate balance of increasing material and transportation costs against the expectation of improved service levels mandated by their key retail customers. This has resulted in several CPG players today focusing on CPFR (Collaborative Planning Forecasting & Replenishment) with their key retail customers as a means to improving perfect order fulfillment, reduced cash-to-cash cycle times and much improved stock availability at the shelf. To gain full advantage of such collaborative initiatives, building the right and an effective supply chain intelligence infrastructure is a must. In addition, with a view to gain information visibility many expensive and technology intensive initiatives like DDSN, RFID, GDS etc. are being undertaken by these CPG players. However, in this race many CPG players are still to establish an effective means to determine the success of these initiatives. Unless this is done, the full potential of such investments cannot be realized. Pg 2

3 Such customer oriented process and technology initiatives are also forcing these CPG players to realign supply chain performance measurement strategies. Traditionally, CPG supply chains have been measured on more internal focused operational metrics like manufacturing capacity utilization, days of inventory, warehouse throughput etc.. While such metrics have helped CPG players to derive internal efficiencies, whether these metrics were aligned to their overall strategic objectives or whether these metrics helped them to win their customers in the marketplace, one is not sure. The focus of this paper is to present various aspects of defining & building the Metrics framework and how this can be effectively leveraged to help drive supply chain performance. A. Building the supply chain metrics framework Through this process, a balanced set of metrics aligned by various supply chain functional areas like Demand Planning, Management, Warehouse Management etc. to meet various decision making requirements from a strategic, tactical and operational standpoint are identified as a first step. These are then tightly coupled to the underlying process & overall business strategy and the roles responsible for executing these processes. A more detailed explanation on this process is provided in the following sections. 1. Establish the right Metrics to measure the supply chain. The first step in building a metric framework is to understand the key characteristics of a metric. Some of these are: Reliability - This refers to the consistency of the metric to measure a given process. As long as the circumstances governing the process do not change drastically, the metric should return a fairly consistent value. A large diversified manufacturer uses Cost of Goods Sold (COGS) as the basis for calculating Inventory Turns. Since the manufacturer had a substantial import content for its raw materials, exchange rate fluctuations led to sharp variations in COGS though overall sales was largely constant. Such variations in COGS led to similar variations in Inventory Turns though Sales & Inventory value the 2 key determinants of inventory turns were constant. Under the circumstances, using COGS as a basis for computing inventory turns made inventory turns an unreliable metric subject to a large amount of random variation. Validity - A valid metric is one that actually measures the concept we think it is measuring. Pg 3

4 Many CPG Manufacturers today are focusing on making their demand planning and fulfillment processes more agile and responsive through a process of greater collaboration among Sales, Supply Chain and Manufacturing. One metric for measuring how effective this collaboration process is working could be the number of expedited work orders. While this metric may suggest inefficient planning in the first place, from a collaboration point of view it can be indicative of a more healthy process of revising supply plans to reflect actual demand changes in the markets. The relevance of metric thus depends on the context. If historically, work orders had been expedited in the CPG supply chain as a result of frequent machine breakdowns or delayed raw material shipments from suppliers, then the metric is not a valid indicator of improved planning process efficiency. However, with a status quo on machine breakdown and supplier delivery times, expedited work orders was a valid indicator of an attempt to revise supply schedules to meet dynamic demand changes. Practicality - In addition to reliability and validity, good metrics must be practical such that the required data can be retrieved with reasonable effort and cost. Salience - They also need to be salient such that the concerned functions/people will relate to the information provided by the metrics and can take meaningful action based on such information. In the above example, if expediting work orders is established as a positive fall-out of the need to make the supply chain more responsive, the Production Engineering function needs to proactively implement practices and tools that allow production line-changeovers at minimal time and cost as soon as they find an increasing trend of expediting of work orders. In addition to the above defined characteristics of a good metric, some additional considerations about metrics highlighted below would help building an effected metrics framework: metrics are most useful when they are embedded in a model that represents a business process the process performance insight that the metrics provide determines the overall criticality of the metric & hence its classification into strategic, tactical or operational metric Constantly monitor and modify the metrics to suit the current business context Assign metrics to various roles that have process execution, monitoring and tracking responsibility. This also helps design information delivery systems based on data and information flow requirement across the organization. Pg 4

5 Identify interdependencies of metrics and an metrics hierarchy which defines how a metric is impacted by other lower level metrics and how this metrics influences other higher level metrics 2. Link metrics to the overall strategic objectives. This again involves multiple steps : Determine the strategic objectives under which you would like to evaluate your supply chain typically these strategic objectives could represent the effectiveness of the supply chain to meet customer requirements or it could also be indicative of the intrinsic strength of the supply chain processes to achieve desired cost objectives. These strategic objectives are an outcome of the overall strategic intent of the enterprise. Under each of these strategic objectives, build related supply chain metrics hierarchy starting with high level metrics suggestive of the overall health of the supply chain to mid level and lower level metrics that are more tactical or operational in nature in terms of impact and review. Some illustrative metrics across different classification has been tabulated below: Strategic Objective Satisfaction Satisfaction Satisfaction Satisfaction Operational Excellence Operational Excellence Typical Metric Class What it signifies Perfect Order Strategic Reliability of the supply chain to meet customer orders in full quantity, in time, meeting specified and agreed quality standards and with complete accuracy on documentation Manufacturing Adherence Schedule Tactical Ability of Manufacturing to supply as per planned manufacturing schedule and hence meeting desired inventory levels for a make to stock item or meeting desired customer delivery schedules for make to order items Machine Downtime Operational Loss of manufacturing capacity due to various reasons like machine breakdown, planned preventive maintenance or stock-out of input materials Supplier delivery schedule Operational adherence Ability of Vendor to supply as per planned supply schedule Cash-to-cash cycle time Strategic Time taken between cash spent to purchase raw material to the time taken to realize cash on sales to customers. This consists of o Days of Inventory o Accounts Receivable in Days o Accounts Payable in Days Total Supply Chain Cost Strategic Total costs incurred in the supply chain including Warehousing, logistics, purchasing, planning, manpower costs etc. Table 1: Metrics Classification Create the detailed metrics framework. This is a multi-step process through which an exhaustive set of related metrics is created. This process involves Associating a set of metrics with each supply chain process. Mapping the metric to the role that is directly accountable and responsible for the measurement and performance of the metric. Certain roles may need to be identified for the metric from a performance review standpoint. Pg 5

6 Identifying the importance of the metric based on the information and process health insight it conveys (Strategic, Tactical or Operational) Build high level interdependencies of metrics based on common knowledge and understanding of basic processes Identify the various hierarchies for which data will need to be obtained to enable a comprehensive view of the metric Determine how the metrics will need to be computed Determine also the frequency at which the metrics will be measured this will also be governed by the granularity of available data and the costbenefits associated with a certain measurement frequency. The table below provides an illustrative metrics framework for the Inventory Management process Metrics Inventory Turnover ratio Inventory Carrying Cost Finished Goods Inventory Performance Attribute Asset Utilization Class Measurement Hierarchy Strategic Product (Average Total inventory value/monthly sales value)*12 Formula UOM Freq. Role Metrics influenced Metrics influenced by No. Monthly Supply Planner, Cash-tocash cycle Accuracy, Forecast Supply Chain Head time Manufacturing Schedule Adherence, Sales Returns Cost Tactical Product Inventory Value Monthly Financial Analyst, Total value*cost of Supply Chain Supply capital Head Chain Costs Asset Operational Product, FG inventory Value% Daily Supply Planner, Days of Forecast Utilization Geography value and % of Sales Manager Inventory Accuracy, total inventory value Inventory Cost Adjustment Tactical Product, Geography Value Physical Value,% Monthly Supply Planner, Total Inventory Warehouse Supply value - System manager, Supply Chain inventory Chain head Cost value Inventory Value, Cost of Capital Manufacturing Schedule Adherence, Warehouse Shrinkage Table 2: Metrics Framework Metrics and Metrics framework are highly interdependent and need to be developed in an iterative and concurrent manner. Metrics framework is useless if they are populated with weak metrics metrics which are not reliable, valid, salient or practical. Metrics by themselves are useless unless embedded in a relevant framework that reflects latest changes in critical business processes. The challenge therefore is to jointly optimize the quality of the metrics and the quality of the metrics framework B. Leveraging the supply chain metrics framework In this section, we demonstrate how the metrics framework can be leveraged to enhance supply chain performance 1. Generate Insights using Cause & Effect Guided Analysis Pg 6

7 A Metrics framework will have a collection of relevant metrics that are inter-related to establish a cause and effect relationship which impacts the overall business process represented by the framework. The fundamental requirement for building insightful and effective metrics framework is a comprehensive and in-depth understanding of the underlying processes. This is illustrated using the example of how retailers, CPG manufacturers and their suppliers are joining hands to build consumer driven supply networks. The figure below attempts to build a simplified supply chain collaboration process model. Greater Collaboration across the supply network Increase availability of POS data & Shelf inventory Enhance visibility of Promotion Plans Improve Store-SKU Forecast Accuracy Increase Product Availability at Shelf Reduce Inventory levels Fig 1 - Supply Network Collaboration Process Model This model highlights how enhanced collaboration (Cause) between CPG players and the retailers can lead to improvements in product availability at the shelf (Effect) as well as improvements in inventory turns (Effect). We can now focus on Product Availability at the Shelf as an effect and start building a metrics model using metrics that would describe the various causes influencing Product Availability at the Shelf. As we do this, we would soon realize that Collaboration across the supply network is one of many such causes. Ability of the CPG manufacturer to reliably meet supply schedules or the ability of the retailer backroom to ensure speedy clearance of goods could be causes impacting Product Availability at the Shelf. A more detailed guided analysis path for Product Availability at the Shelf has been depicted in Fig 2. Pg 7

8 Product Availability at the Shelf Store Inventory Availability Backroom Clearance time CPG Manufacturer OTIF Delivery Performance Forecast Accuracy Ordering Efficiency Documentati on Accuracy Staff Productivity Order Fulfillment Time Quality Adherence POS & Shelf Inventory Availability Promotion Plan Visibility Order Quantity vs. Demand Ordering Frequency Transit Time variance Manufacturi ng Schedule Adherence Warehouse Productivity Collaboratio n Process Adoption Fig 2 A Guided Analysis path for Product Availability at Shelf This guided analysis helps determine the root causes impacting product availability at the shelf. As is evident above, Collaboration process between the CPG manufacturer and the retailer is only one of the many processes impact product availability. This model also highlights the need to collect many other important metrics like Forecast Accuracy, Order Fulfillment times, Manufacturing Schedule Adherence etc. In the absence of a good model, we would only have data regarding Product Availability at Shelf and not of the related metrics, severely limiting the ability to use the end result data in any constructive manner. Each enterprise is expected to have a metrics model that is unique in terms of number and types of causes, hierarchy of causes and criticality of causes to the end effect. Some of the key points to be considered while defining the metrics model can be summarized as follows: No metrics model is likely to identify all relevant variables however, enterprises need to make a conscious effort to build a metrics model that is comprehensive while providing an in-depth cause and effect analysis. This can be best achieved by deploying a cross-functional team that has a strong understanding of underlying processes. More process-oriented companies also support such teams with advanced statistical and quality tools to help them determine highest impact causes. Most importantly, such a team should have a stake in terms of accountability and responsibility for the metrics being measured and the resultant actions emanating from these metrics. Pg 8

9 Despite use of advanced statistics, some element of judgment and approximation on behalf of the team developing the metrics model is unavoidable. Given this imperfection, the attempt therefore should be to identify the most relevant & highest impact causes keeping in mind that each additional metric in the model requires investment of time, effort and money to collect the right data and to ensure its quality over time. Like metrics, models tend to be highly contextual in nature. They are very much dependent on the team of people developing them their knowledge of underlying processes and causes, goals for measurement and their vision of how a supply chain performance measurement system should integrate with the overall enterprise performance management objectives. Besides people, another important factor which constrains the ultimate design of the metrics model is data in terms of availability, accuracy and cost of obtaining this data. In the above example, getting accurate and real-time data on product availability is itself likely to pose a major challenge. This in turn, would decide how the metrics model would ultimately shape up. Metric models cannot be static. Depending on the nature of business changes, changes in process maturity as well as changes in the vision of the enterprise, metrics models will need to be refined to align with the latest enterprise objectives and needs. As the IT landscape of the enterprise changes, some metrics which might have been consciously ignored because of data availability reasons may now need to be accommodated in the model to make the model more robust and relevant. 2. Data Analysis across various dimensions While a metrics model provides the various metrics or attributes for analysis, another dimension of analysis is the hierarchy of data. Typically, data hierarchy allows one to move up or down on a particular attribute. F O R E C A S T A C C U R A C Y Product Hierarchy Geography Hierarchy Role Hierarchy Time Hierarchy Division Categor Brand SKU y Country Region State City Head of Sales Regional Manager Area Manager Territory Executive Yearly Quarterl Monthly Weekly y Fig 3 An illustrative data hierarchy In the previous example, while forecast accuracy is a metric for measuring the demand planning process and Promotion Plan visibility one of the causes impacting forecast accuracy, the analysis would be more meaningful if we could also conclude on the performance level of forecast accuracy across products, across geographies and across Pg 9

10 roles participating in the demand planning process. Knowing this would help arrive at the exact location of the root cause. Using such data hierarchy, one would also find that a particular metric can have multiple causes related to a specific data hierarchy. For example, forecast accuracy at a Country level may be the key cause for poor performance of promotions. However, going down the geography dimension to a particular region, one may find that at a regional level, it is not forecast accuracy that is the key cause but it is supplier compliance to schedule that has led to delayed supplies and hence poor promotion performance. This helps us to focus on regional supplier compliance issues while addressing forecast accuracy at a national level. 3. Quantify financial impact of supply chain metrics using Scorecards. This is the process be which typical supply chain metrics are linked to financial KPIs. For example, Cash-to-cash cycle time metrics model can be linked to Return on Assets. Similarly, Total Supply Chain Cost metrics model can be linked to Net Margin through Cost of Goods Sold. Establishing the link with financial measures helps quantify the performance of Supply Chain as well understand its full impact on the enterprise s top line and bottom-line. This is an important step. When supply chain metrics like Perfect Order or Cash-to-cash cycle time are linked to financial measures like Profitability or Return on Assets or Investments, they provide a completely new focus and rigor to how these processes are managed. A sample Supply Chain scorecard is produced below : Typical Metric Performance Attribute Strategic Objective Benchmark Baseline Entitlement Target Actual Benefits from Improvement Perfect Order Reliability Satisfaction 75% 90% 80% 60% $50M savings in Lost Sales revenue Order Fulfillment Time Responsiveness Satisfaction 3 days 2 days 2 days 3 days $10M additional revenue from surge orders Forecast Accuracy Cash-to-cash cycle time Total Supply Cost Chain Cost as a % of COGS Process Improvement VMI as % of Process total Asset Utilization Operational Excellence Operational Excellence Operational Excellence Collaboration Efficiency 85% 95% 90% 75% Key enabler to Satisfaction and Operational Excellence 90 days 75 days 85 days 105 $70M reduction days in working capital; $10M savings in interest cost 15% 10% 10% 18% $45M reduction in Direct Cost; $20M reduction in Indirect Cost 50% 65% 50% 15% $15M reduction in working capital Table 3: Supply Chain Scorecard Pg 10

11 Using such scorecards help enterprises determine priorities for investments to improve processes and related technology. Such a scorecard also helps establish a standardized single version of truth on supply chain performance that is quantified and can be easily understood by all entities of the organization. 4. Review the Supply Chain Scorecard in the Sales & Operations Planning process. While metrics reflect the overall health of the supply chain and its various functions, they need to be supported with a process mechanism which enables a joint review and formalization of corrective plans from a cross-functional perspective. With the emergence of consumer driven supply networks, such a mechanism may also involve multiple players like retailers, CPG manufacturers, third-party logistics service providers and vendors in the review process. Some key best practices for implementing effective S&OP processes include : Establish a prescheduled meeting with a well-defined agenda involving Retailer, CPG Manufacturer & any other critical players in the supply network along with their cross-functional teams Articulate and quantify the performance for the supply network as a whole which in turn leads to performance measures for individual players of the supply network and also helps speedy determination of causes of failures Need to have representation by key decision makers including Executive involvement that can take deterministic action and unite multi-functional objectives Adequate groundwork on performance review and failure analysis needs to be undertaken at a functional level to the supply network performance review in the S&OP meeting so that the S&OP meeting is more of a confirmation only with no surprises Benchmark supply network performance with the best in the industry and across industries use these benchmarks to continually raise the bar on performance On the other hand, one of the key challenges to enable an effective S&OP is to have an efficient IT capability to aggregate and structure enormous amounts of supply chain information & data originating from disparate IT systems such that it enables an overall view of the supply network. Delivering the Metrics framework This is the domain of Supply Chain analytics. Supply Chain analytics is the process of extracting, transforming and presenting supply chain information on a common information platform in various presentation formats like dashboards, reports, alerts to fulfilling the diverse information needs of operational managers, senior executives as well as remote users or business partners. At the core of the solution is a Data Warehouse being fed supply chain information from various operational transactional systems like ERP, CRM and other data sources supporting the various supply chain process. The solution maybe powered by Pg 11

12 sophisticated analytical tools for smart reporting, generation of exception alerts or for predictive modeling and scenario planning. The front end of the Supply Chain Analytics solution needs to be speedy, easily accessible and user friendly. One of the best ways to present & deploy supply chain metrics to the various levels of the enterprise is to use Dashboards that are bolted on the data warehouse. They provide the flexibility to customize presentation of data for various supply chain metrics in a variety of formats like graphs, dial indicators etc. all in a single screen. Each screen, metrics displayed on the screen and drill feature can be customized to meet the varying requirements of the users. An operational visualization of the Supply Chain Analytics solution is that of executives and managers at different levels of the supply chain function making day-to-day operational, tactical and strategic decisions through supply chain metrics embedded into scorecards and presented in the dashboards. For example, the customer service manager company can open his/her Dashboard first thing in the morning and immediately have a quick understanding of customer-wise sales and service levels achieved across different warehouse for the previous day. He can use this snapshot to build certain supply priorities for his customer and their orders. In most cases, he would also be supported with Alerts guiding him to establish priorities for the day. He can also undertake an investigation across various product & location dimensions to pinpoint the warehouse and the product where there has been a supply disruption. On the other hand, the Head of Supply Chain can see the 5 most crucial supply chain KPIs defined by him using some standard daily or weekly reports. If he decides to investigate any of the metrics further, he can drilldown into the metric as well as slice and dice it by different dimensions of the data. For senior executives, the best practice is to provide a guided analysis path using metrics framework, thereby restricting the number of metrics, the number of drill paths and the number of dimensions by which he can slice and dice the metric. The metrics, primarily strategic and tactical, are tailored for him and secured so no lower level associates would be able to view them. Closing Comments Enterprise Performance Management (EPM) describes a process-centered approach to business decision-making. It is meant to improve a business s ability to gain insight and manage its performance at all levels using a metrics framework combining stakeholders, managers, and employees within an integrated management environment. EPM should be a business critical process driven by metrics, supported by business intelligence and executed by people. Given the increasingly competitive and constantly changing economy, tapping into this critical asset is the key to sustaining competitive advantage. Pg 12

13 Bibliography: 1. Metrics and Models for the Evaluation of Supply Chain Integration by Jonathan A. Morell, Ph.D. 2. Supply Chain Clarity Issue 5 Feb 2005 the newsletter of Supply Chain Analytics Ltd 3. The Intelligent Supply Chain : Taking Enterprise to the Next Level by John Hughes 4. Enterprise Performance Management : Article by Data Management Group About the Authors: Tejas Faldu (tejas_faldu@infosys.com) is a Senior Consultant with Retail and CPG practice of Infosys. He has over 10 years of domain experience in various areas of supply chain management in the CPG industry. He is currently involved in developing business led IT solutions for the Retail & CPG industry. Srikanth Krishna (srikanth_krishna@infosys.com) is a Senior Project Manager with Retail and CPG practice of Infosys with 9 years of experience in this Industry. He has worked with large global organizations to develop and implement solutions for Enterprise-wide initiatives on Business Intelligence, Data warehouse & Intranet Portals with Dashboards About Infosys Technologies Ltd. (NASDAQ: INFY) Infosys (NASDAQ:INFY) defines, designs and delivers IT enabled business solutions. These provide you with strategic differentiation and operational superiority, thereby increasing your competitiveness. Each solution is delivered with the industry-benchmark Infosys Predictability that gives you peace of mind. With Infosys, you are assured of a transparent business partner, business-it alignment with flexibility, world-class processes, speed of execution and the power to stretch your IT budget by leveraging the Global Delivery Model that Infosys pioneered. Infosys Retail and CPG Practice The Retail & CPG Practice provides business solutions to Blue chip clients enabling them to become more competitive. Our client base spans across department stores to apparel designers and, specialty retailers, distributors and CPG manufacturers. The 2000 strong practice provides services that include business process conceptualization, process engineering, package selection and implementation. We are an UCCnet certified company and also an active member of ARTS (Association for Retail Technology Standards). Pg 13