Master Data are the fundamental business data in the company, typically long-lived and used across multiple applications, inherently including the subset of Master Reference Data.
Master Data, including Reference Data, is not to be considered “Metadata”. Metadata is data about data. It describes data content but it is NOT the content. There is no formal and universal definition of how deep to take the definition from a content perspective. Only those metadata whose management will
bring more value to the enterprise than the cost of the labor needed to create and maintain it should be managed and integrated formally. For example, “Master Data” definitions (customers, suppliers, products, organization, etc.) which will be the most ever-present and shared data across an enterprise will be most critical.
Master Reference Data are used to understand, navigate and query information based on or related to the Core Master Data from various business level and/or user level perspectives.
Value Proposition of MDM:
- Enables 360º view of the enterprise and corporate performance management improvement.
- Provides competitive advantage by enabling customer behavior insight & predictive modeling.
- Improves the forecast management through more effective logistics and inventory control.
- Reduces cost of manual master data reconciliation & alignment efforts, error fixing, etc.
- Reduces the data redundancy cost by consolidating & eliminating duplicate masters (DB/apps).
- Reduces application development & maintenance costs, by having clear master data interfaces.
- Improves the decision quality by securing reliable, high quality master data “just in time”.
- Improves the availability of key data – speed of access, data timeliness, common (user) support.
- Eliminates/prevents redundant & non-coordinated master data activities within the company.
Executing components of the MDM Strategy
- Data Governance
- Key roles
- Desire to roll-out data governance to other areas of the business
- Data Quality
- Incorporating DQ approach into EDW instance consolidation project
- Process Improvement
Having spent so much time in the Entertainment and Hospitality industry, I see everywhere the key systems being CRM and Revenue Management systems. Yield Management is a market segmentation strategy to capture the consumer surplus, widely used in the Entertainment and Hospitality industry. Airlines, hotels, theme parks, car rentals, cruise ships, broadcasting (TV, radio), and Utilities (telecom, electricity), etc. always used segmentation and peak-load pricing. But with the advent in computing power of systems this technique is now able to do excessive and data intensive calculations with linear programming to optimize revenue. Just like airlines realized long ago that a flight flying with an empty seat is forgone revenue, many companies use historical data to optimize differential revenue gains. American Airlines had added $1.4B additional revenue over three-year period in the early 1990s. Hertz added 1-5% revenue annually and so did companies like Marriott Hotels, Royal Caribbean Cruise Line, etc. Computer algorithms can use variables like time of purchase/usage (advance/spot purchase, day-of-week/season), purchase restrictions (cancellation options, minimum term, Saturday night stay), purchase volume (individual vs. group), and duration of usage (single night/weekly rate) to get the right price to the right customer. Next time you are stuck an airport and an airlines mentioned over-booked flight, know that some computer algorithm was doing some marginal analysis based on capacity of flight, typical cancellation rate, revenue from booking seat and cost of denied boarding (these days with the pressure airline are also adding costs like loss of goodwill along with the free flights they give you).
Other analytic systems within an organization depending upon what the goals of a business or its units are.
- Monte Carlo Simulation – This technique uses data to establish a pattern between a domain of possible inputs. The calculations generate inputs randomly from a probability distribution over the domain, perform computations on these and aggregate the results. This tries to minimize reinventing the wheel by reuse of research results. In the Pharmaceutical industry, for example, during the Discovery phase of their value chain, data mining is used to search contextual information based on secondary relationships,
- Regression Analysis – This is a set of techniques used for modeling and analyzing several variables and establish a relationship between a dependent variable and one or more independent variables. It is widely used in Marketing and Sales part of a company’s value chain since it can provide deep insight on customer behavior and provide an enhanced decision-making for future customer interactions. Companies are using industrialized analytics which uses closed-loop promotion and data mining to optimize marketing budget allocation (i.e. optimizes marketing channel and product mix). The closed-loop-promotion ensures feedback of mission-critical data to marketing and provides on-demand access to marketing decision-support. This helps creates an optimal marketing mix (as Jerome McCarthy called this the 4P’s – product, price, place, and promotion). It started with improvements in systems for statistical analysis that helped understand why certain things were happening in the business environment. These were heavily used to do web traffic analyses during the emergence of eCommerce. Then these systems were extended to forecasting and extrapolation to see what would happen if the trends continued. Then the era of predictive modeling built systems what could try to predict what will happen next, based on empirical data and heuristics. Of course all these systems are used to help with optimization of spend and maximize the revenue. As an example, within the financial services industry, specifically Banks, such models are used to predict Retention & Loyalty, perform Portfolio Analytics, help with Fraud (Transaction/ Payment) and Anti Money Laundering efforts. They help answer questions like Bank Servicing – Which customers/ segments are at risk? Which ones are profitable? How do I retain them? How do I win back customers? They can also help with Loyalty Programs – What drives loyalty for my customer base? How do I design an effective loyalty program? They are used for Cross Selling and Upselling – Which deposit account holders would be interested in an auto loan? Can we sell insurance along with auto loan? Which products can be sold on credit card welcome call? Which ‘Gold’ customers would increase their spend if we upgrade them to ‘Platinum’? Can I identify customer requiring short term loan? It helps with Campaign Management with segmentation modeling, profitability analysis, retention campaigns, win-back campaigns, etc.
- Neural Network Analysis – Apart from the obvious impactions in the study of real biological neurons, this technique is used in areas like distribution & logistics, sociology, economics, etc. Any company that produces a product is chartered with getting that product to its customers (whether that’s the final customer, retailer, wholesaler, etc.) in the least expensive manner. They are constantly trying to reduce total inventory while maintaining service levels of supply at each warehouse or distribution center location. Neural network models help with optimizing with the variables involved: number of warehouses to have, location of each warehouse, size of each warehouse, allocation of products to the different warehouses, allocation of customers to each warehouse, etc. The objective is to balance service level of supply against production/ purchasing costs, inventory carrying costs, facility costs (storage, handling and fixed costs), and transportation costs.
Sure a picture is worth a thousand words. But didn’t the dude in Matrix see numbers and know the stories behind this? Well, only in movies. According to research, Data visualization is so powerful because the human visual cortex converts objects into information quickly. As we continue the journey of Data – Information – Knowledge – Wisdom, the feedback loop of models and visualization to see patterns is key.
As Big Data grows, it’s clear that the technology to gather and store data far EXCEEDS the ability to Analyze it. However, not all visualizations are actually that helpful. You may be all too familiar with lifeless bar graphs, or line graphs made with software defaults and couched in a slideshow presentation or lengthy document. The best data visualizations are ones that expose something new about the underlying patterns and relationships contained within the data. Understanding those relationships — and being able to observe them — is key to good decision making.
- Pizza and Cola sell together more often than any other combo – is there a cross-marketing opportunity?
- Does Plant and Clay Pot sales IMPLY sales of Soil?
- Milk sells well with everything – people probably come here specifically to buy it. Should we raise prices since less price elasticity?
- What is the one item you want to have in your store in case of a hurricane?
- Does buying any kind of pepper also denote sales of banana?.
- Does buying any kind of pepper also denote sales of banana?.
- Which customers are most likely not to have an accident?
An important distinction lies between visualization for exploring and visualization for explaining. Exploring data is all about statistical acumen and understanding the nature of what the data represents in your enterprise. Visualization tools are an aid but they have been around for eons. Once you have explored, you will almost always find less than a handful of factors stand out and need explanation. Your presentation should not be about fancy graphs but the right power point / keynote /video storyline for your audience. It seldom needs voluptuous graphs … if you are trying to describe more than this handful of points, then you are already lost in your quest.
The key is use the right Visualization for the right Data at the right Time. I found this chart very helpful to decide the decision tree for which types of visualizations to use for different scenarios:
There are so many tools to do this kind of analyzes:
- Qlik, SAP, SAS, and Tableau Software deliver the latest table stakes in visual discovery: storyboard capabilities.
- Google Fusion Tables: Bust your data out of its silo and combine it with other data on the web. Collaborate, visualize and share
- Datawrapper: An open source tool helping anyone to create simple, correct and embeddable charts in minutes
- Infogram: Infogr.am is user-friendly interface to help develop creative, interactive infographics
- Piktochart: Piktochart is a simple WYSIWYG editor to help develop and design charts and infographics
Visualization for explaining is best when it is cleanest. Here, the ability to pare down the information to its simplest form — to strip away the noise entirely — will increase the efficiency with which a decision maker can understand it. As big data becomes bigger, and more companies deal with complex datasets with dozens of variables, data visualization will become even more important.
So last week when I wrote about Predictive Analytics, I got responses from folks saying, “The value from such areas is clearly there. But the challenge is which technique to use and the ever-sliding sword of showing ROI so there is buy in for these analyses”.
In framing the Analytics problem – we need to balance data, SME knowledge, and performance. One of the things I have noticed in my work is when the analysts build models the real skill in creating effective analytic model is knowing which models and algorithms to use. They can use different techniques: neural networks, decision trees, linear regression, naïve Bayes, etc. But these days many analytic workbenches now automatically apply multiple models to a problem to find the combination that works best. One needs to explore different paths – they look at the problem from different perspectives. When these algorithms are combined there is resulting synergy. Once the modeling data sets were finalized, the largest incremental gain was not achieved by fine tuning the training parameters of an individual algorithm, but by combining predictions from multiple algorithms.
So with the myriad tools and techniques that exist, the way to approach this is to ask the questions that are really important for what the company is trying to solve:
- Strategic Customer Questions
- Who are the most/least profitable customers?
- Who are the most/least satisfied customers?
- What is fastest/slowest customer segment?
- What are the reasons for customer attrition?
- What are the costs of customer transactions?
- Strategic Product Questions
- What are our most/least profitable products?
- What are our production costs & how can we lower them?
- What is our cycle time & how can we lower it?
- Strategic Employee Questions
- Who are the most productive salespeople, employee?
- Which managers have the highest retention rates? What do they do?
- What is the cost of turnover?
- Strategic Financial Questions
- How accurate are the financial forecasts?
- How much financial data is used to answer business decisions?
- What impacts the demand of our product?
- What items are affecting our margins the most?
Based on this one has to look at some of the following techniques:
- Classification – predicting an item class, “Decision Tree”
- Association – what occurs together, “Market Basket”
- Estimation and Time Series – predicting a continuous value
- Web and Text Mining – extracting information from unstructured data
- Clustering – finding natural clusters or groups in data
- Deviation Detection – finding changes or outliers
- Link Analysis – finding relationships
These days it’s tough to walk out of any meeting with Business or IT organizations without touching the topic of Big Data or Analytics. Lot of people struggle with what this is – everyone BELIEVES that it can help if done rightly. But what is it?
The way as I look it: As organizations mature on their Business Intelligence capability, the questions they ask mature too. It’s not about only looking at what the data tells about problems you need to solve. But can data tell you to THINK OF NEW PROBLEMS that you can solve. Things you didn’t know. THINK of something different. Organizations are faced with ever increasing business challenges: Driving new sources of growth, Cost management and cash conservation, Increased business complexity and the need for operational excellence, or Business restructuring in an increasingly global business environment. Ubiquitous computing and technology capabilities have increased dramatically the volume of data at companies’ disposal, yet there remains little in the way of actionable insights (Big Data). Companies need timely, in-depth actionable insights if they are to remain competitive globally to effect a “whole business” approach to big data analytics to deliver business results. Analytics-driven optimization of key business processes
- Staking out distinctive market strategy (CRM Strategy and Loyalty programs)
- Finding the best customers, and charging them the right price (Revenue Management )
- Minimizing inventory and maximizing availability in supply chains (Inventory Optimization)
- Understanding and managing financial performance (Forecasting)
Business Intelligence technologies are deductive in nature validating the hypotheses of the business problems you want to solve. Examples:
- Product shortage by market
- Vendor spend by category
- Brand health by market
- Periodic trend analysis
- Periodic P&L and Financial Reports
Predictive Analytics is Inductive in nature – pull out meaningful relationships and patterns and tells you of different things that might be addressing the same or new problems. Example:
- Business Mix Optimization (Product, Geography, etc.)
- Price sensitivity by consumer segment
- Customer Behavior Modeling
- Performance/profitability analysis
As an example, NETFLIX, a US movie delivery company, asked engineers and scientists around the world to solve what might have seemed like a simple problem: improve Netflix’s ability to predict what movies users would like by a modest 10%. From $5 million revenue in 1999 reached $4.3 billion revenue in 2013 as a result of becoming an analytics competitor. By analyzing customer behavior and buying patterns created a recommendation engine which optimizes both customer tastes and inventory condition.
As another example, Analyzing Love: Data Mining on Match.com in AllAnalytics, Match.com, online dating service, tries to predict the likelihood of attractions between people. 95% of relationship can be predicted by analyzing as few as 10 characteristics in each profile. The find things like:
- Members with accounts on Twitter, which only allows for messages of no more than 140 characters, have shorter relationships.
- People identifying themselves as Republicans are more willing to connect with Democrats than the reverse
In this day and age there is an assumed maturity in the way initiatives within a business are sourced and out-sourced. When it comes to IT applications and their development and maintenance, there are 4 possible scenarios that companies deal with:
- Insource - Maintain control internally (usually for reasons of intellectual property, privacy, or strategic responsiveness)
- Staff Augmentation - Save money while maintaining responsibility for application support and maintenance activities
- Co-source- Leverage external cost structure benefits and expertise while maintaining an appropriate level of control
- Outsource – Delegate IT (or selected functions therein) to an external organization for which it is a core competency
With this industry evolved over the years, the rationale for IT outsourcing decisions has shifted from cost being the sole consideration to include a number of strategic factors. No doubt cost is still top of the mind, especially with this economy. But a lot of other considerations are in play:
- Strategic Importance
- Relative impact of a service area on the company’s revenues and overall profitability
- How strategic is the function to my organization today? How does it fit
into our future plans?
- Current Capability
- Relative strength of a service area’s technical & business know-how, processes, and tools
- What are the capabilities of the function? How do those capabilities compare to our requirements, and to our peers?
- Perceived Value / Cost
- Perceived value of a service area relative to the costs incurred
- What is the function’s capacity to adapt and change?
- Ownership Preference
- Relative preference of management to own, share, or transfer out IT assets based on company beliefs, values, and sourcing experience
- How easily can the function be transitioned to another sourcing strategy?
Business Quarterly indicates 75% of US executives considered financial motivations as secondary to other strategic objectives when outsourcing. Business Week reports, “The really smart business owners have figured out how to use outsourcing as a strategic tool instead of simply looking for savings.” CIO magazine reveals strategic value rivals cost reductions for outsourcing motivations.
Based on some reports by The Outsourcing Institute the top reasons for outsourcing look as below:
No matter what the goals, the key success factors of outsourcing are always:
- Be clear about objectives– cost, process improvement, and the ability to focus on the core business are the most common
- Incorporate business outcomes as a performance measure from the outset of the arrangement
- Look beyond price and promises of cost reductions for an outsourcing provider that brings a wide set of skills and strengths, and a long-term track record of delivering results
- Give as much attention to performance measurement and the quality of your relationship with your provider as you do to the contract
- Use active governance to manage the outsourcing relationship for maximum performance
- Task talented executives with optimizing outsourcing arrangements
Ever since Jerry Maguire blurted this out, people have been using this as a corporate euphemism for ROI/ Business case.
One of the critical roles for any organization is to manage the value achievement of the initiatives they pursue. They need to ensure sponsor and executive ownership of the business case. The business case allows the stakeholders in IT projects to jointly address their key concerns with project investments:
Business cases highlight the initiatives that create the greatest value, support decision- making, and help track program performance. It is good to define the business case early and plan on many iterations since it:
- Demonstrates how a major investment creates value
- Includes both quantitative and qualitative rationale
- Supports business decisions by weighing choices or options
- Creates a way to track performance and measure success after a decision has been made
- Gains alignment and management consensus for a project
In some organizations, the term ‘Business case’ may also be referred to as
- Cost/benefit analysis
- ROI analysis
- Feasibility study
- Capital funding request
- Case for action
- Once the team has understood the importance of having a business case to guide the investment decisions of the initiatives, there is debate on what level of detail should it have. There are many approaches to building out a business case and the main elements are
- Benefit models
- Cost models
- Cash flow models
- Assumptions (timing, dependencies)
- Sensitivity Analysis
- Qualitative Factors Analysis (non-financial benefits, risks)
The financial models can be Top-Down (more high level and helps form an initial hypothesis wider ranges to reflect uncertainty) or Bottoms-Up (more quantitative and time spent on thorough data collection and analyses). But the key point is that you need to build the business case with ranges and confidence levels. Once the numbers were compelling, the ranges could change but they would not change the decision.
At a BPM event recently in Orlando, I was chatting with a colleague about IT and the BPM responsibility. This guy is the SVP of IT operations and handles Infrastructure for his company. When someone asked him who from the business was responsible for the BPM aspects in his firm from the business side, his response was “We in IT are actually responsible for the BPM aspects and optimization therein.” Another guys goes, “The only real applications the business is concerned about is e-mail”
That set me thinking about IT Service Management, etc. Having spent some time doing ITIL work, I am familiar with the concept of IT service management, which involves moving:
- Multiple points of contact with the business
- Service defined and measured in technical terms (if at all)
- Work driven by technology
- Organized to support systems
- Managed relationships established with customers
- Service defined, measured and reported on in business terms
- Work driven by service requirement
- Organized to deliver service
So ITSM is all about better service at lower cost. But the challenges with a full blown ITIL deployment is that ITIL is far too generic for an organization to implement at a fast pace, in totality. Process reengineering and change management are always required and are rarely considered. Some practitioners have said that it complements other IT management methodologies like CMMI, etc. But the way I look at this is that CMM focuses on improving and appraises the maturity of application development. ITIL is focused on best practices around IT Operations and Services. This kind of demarcation:
The ITIL v2 broke these Operations into Service Support (ensuring that the customer has access to appropriate services to support business functions) and Service Delivery (IT services are provided as agreed between the Service Provider and the Customer).
But the key to achieving good IT service management even at a small scale is by using the following guiding principles:
- Business Relationship Management: Ongoing liaison and relationship building with Client community. Maintain an understanding of the business and IT requirements.
- Service Delivery Management: Understand the IT Services provided and the businesses reliance on these Services. Carry out the appropriate business liaison and escalation for Service issues.
- Service Performance Review: Formally review service performance against agreed upon SLAs. And good luck with that J
- Service Level Agreement Management: Maintain service definitions and assess implications of any changes
- Service Enhancement Request: Receive and shape requests for new/enhanced services
Achieving supply chain excellence is complex and challenging, but success in achieving supply-chain driven competitive advantage enables superior customer service, profitable revenue for growth and significant increase in shareholder value. Inventory Management is the conductor of the symphony for Retail Supply Chain execution. It is critical for customer service since Inventory management is what initiates all merchandise movement and controls the timing within the supply chain
- Supply chain assets and inventory usually comprise at least half of all non-store based assets
- Supply chain activities typically account for as much as 40 – 70% of operating costs (including procurement and markdowns)
Some of the statements from retailers across all kinds of products:
- “Assisted Inventory Management (AIM) helped us exceed our inventory-turn goal, making us the leader among national drugstore chains in this important productivity measure. We achieved inventory turns of 5.0 times for the year, up from 4.6 times in earlier years.” – CVS
- “Positioned among the best in retail, our supply chain helps drive sales, reduce costs and ensure the availability of products our guests most want and need.” – Target
- “We completed the conversion of each of our operating divisions to a common technology platform with greatly enhanced inventory management tools, permitting more sophisticated inventory planning and more precise by-store inventory allocation.” – Saks
The three main components of the Inventory Optimization program address both the process and physical infrastructure of the supply chain.
- Inventory Management Process – this addresses end-to-end inventory management built on two core processes:
- Foundational for continually replenished basic merchandise. Periodic automatic replenishment, long life, stable supply, short lead time to continually meet normal demand
- Highly Variable which is typical of merchandise with high demand spikes due to promotions, fashion, short life and seasonal demand
- Network and Flow Strategy – Network Optimization starts with establishing a vision of alternative flow paths and ends with a full evaluation of end-to-end physical supply chain and a recommended distribution network strategy. One has to assess merchandise flow paths to provide revenue growth, minimize supply chain costs and support overall inventory strategies. Then one has to determine alternative distribution strategies including buildings size and location, transportation strategies, inventory deployment strategies, and benefit based business cases.
- Store Operations – Design and implement a well-defined process for store operations related to receiving, shelf stocking, perpetual inventory accuracy and plan-o-gram maintenance.
- Organization & Labor Planning
- Life Cycle Management
- Shelf Replenishment
- Data Integrity Maintenance
The idea is to push operations from
- Stores Ordering for basic merchandise to Automatic Replenishment Approach which is centrally maintained and helps with enhanced High Performance forecasting and allocation abilities
- Store Reviews ( All replenishment orders to supplement simple forecasting & ordering logic) to Exception Only Reviews. No store review for standard items and examples of exception reviews: items with high inventories, poor service levels etc.
- Limited Standards & Policies (In-stock policies and Service levels) to Standard Policies Across the Supply Chain. This is through reliable & repeatable inventory management processes and uniform service standards based on merchandise goals and category/SKU profitability
Having done a lot of work in the supply chain industry, I am so intrigued by RFID and its potential once the costs go further down. Radio frequency identification (RFID) is a generic term for technologies that use radio waves to automatically identify individual items. RFID technology is not new or complex; it has been around since the early radar systems in the 1940’s. What is new is how manufacturing advancements have reduced costs of implementing RFID systems (particularly tags). These silicon-based electronic identification tags, consisting of a tiny processor, memory, antenna and can be read and written wirelessly and can be made cheap, without a battery. The main components of this technology are:
- Device made up of an electronic circuit and an integrated antenna
- Radio frequency used to transfer data between the tag and the antenna
- Read-only or read / write
- Receives and transmits the electromagnetic waves
- Wireless data transfer
- Receives commands from application software
- Interprets radio waves into digital information
- Provides power supply to passive tags
- Reads / writes data from / to the tags through the reader
- Stores and evaluates obtained data
- Links the transceiver to an applications, e.g. ERP
Of course there has been a major drag in the adoption of this technology. The key challenges have been:
- Not only costs of tags and readers, but the costs of integration of the RFID technology into the IT technology stack – e.g. ERP, etc.
- Lack of worldwide data standards
- Country-specific frequencies allocation
- Vendors are very fragmented
- Tag and data overload – How do we handle the data?
- Read-rate accuracy
- Tag and reader collision – Signals can interfere with each other
- Privacy fears from the tracking provided by this technology
But more and more this technology is coming into mainstream. Especially after Walmart mandating the use of RFIDs in their supply chain management. Walmart believes that they can cut out costs and make their supply chain even more lean with this deployment.
The uses of this technology are of course endless. I was recently reading about the CyberTM Tire from Pirelli Tire Systems that transmits information on road conditions and friction coefficients to the car’s computer. Already some hospitals are using RFIDs to tag patients with wristbands to scan by hospital staff using PDAs or tablet PCs connecting to patients’ data using a WLAN.
And as this become more prevalent there are other uses that are surely ridden with privacy issues. There is much research where people are looking at ways to monitor real time health in individuals. There is a RFID implanted in the human wrist that send signals to the health insurance company at all times. When you wake up in the morning and go for a jog; you arrive at work and an email from the company (always monitoring your vital stats) sits in you inbox, proclaiming a reduced premium for the day. You have breakfast at McDonalds over the weekend. Lo and behold, your premium just went up.