The entire dashboarding process can be mapped along the answers to these three questions:
1. What information?
2. For whom?
3. How to present?
Meta-information is information about information. For those from the information technology (IT) or database profession, this would resonate with the concept of metadata -- data about data.
Dashboard implementation at the level of meta-information requires collecting information about the information required to display through dashboards. On a primary level, this process involves determining the critical business questions that need to be answered through dashboard deployment, and then mapping these questions to the key performance indicators (KPIs) that need to be captured through the dashboards in order to get answers and insight.
The first step in this process is to document all of the KPIs that are currently being captured through regular reports or ad hoc analysis. Usually, such KPIs are well documented and mapped to the various data sources required to generate them. At this stage, it would also be useful to determine if any KPIs are desired by decision makers but for some reason are not currently made available.
Additional KPIs can be identified through interviews and focus groups with decision makers. During these meetings, executives and managers should talk through the steps required to complete their decision making processes and the information required for each step. From these meeting, the most important KPIs will be identified. These KPIs should be displayed on the highest level dashboards. Secondary or rarely accessed information should be accessed through drill-downs from these highest level dashboards.
Another strategy for thorough cataloging of KPIs is to divide them by various divisions within the organization (e.g., Sales, Marketing, Manufacturing, Supply Chain, Customer Service, Human Resources, Finance). It is very likely that at any given time, a dashboarding effort deals with a single division, and therefore, for the most part, the necessary KPIs are relevant to that specific division only. However, eventually the needs of the organization will require that dashboards within one division tap into KPIs from other divisions in order to build an understanding of the big picture. Senior executive dashboards would invariably require consolidation of KPIs from all divisions.
The following are some standard KPIs for various divisions:
- Sales KPIs. Gross and Net Revenue, Unit Sales, Number of Orders, Average Order Value, Pipeline Conversion Rate, Active Forecast, Days Since Last Sale, Revenue per Employee
- Marketing KPIs. Percentage of Promotion Response, Price Elasticity, Unique Web Site Visits, Product Sales Mix
- Supply Chain KPIs. Cycle Time, Stock Quantity, Lost Sales Volume, Return Rate, On-Time Delivery Percentage, Inventory Turn
- Customer Service KPIs. Retention Percentage, Cross-Sold Percentage, Service Level Met Percentage, Handling Time, Call Time, First-Time Resolution Percentage, Referral Percentage, Unresolved Percentage, Case Ageing
- Human Resources KPIs. Headcount, Employee Turnover Ratio, Average Tenure Length, Hire Cycle Time, Skill Level
- Finance KPIs. Revenue, Gross Profit, Gross Margin Percentage, Net Income, Accounts Receivable, Cash Flow, Asset Turnover Ratio, Cash Flow
- Manufacturing KPIs. Labor as a Percentage of Cost, Downtime Percentage, Variance from Plan, Time from Order to Shipment, Time on Floor to be Packed
This article is the answer to a question from a Dashboards 101 reader. As an article featured in Dashboards 101, the reader who asked this question has won a free signed copy of the dashboarding book Enterprise Dashboards: Design and Best Practices for IT by Shadan Malik (John Wiley & Sons, 2005).
So email your Questions, Comments, Concerns to firstname.lastname@example.org and you could be our next winner!
Defining the Key Performance Indicators
The process of defining the KPIs for dashboards is no different than doing the same for building a reporting infrastructure (or business intelligence system). To perform this process with full rigor and detail, an experienced information analyst is required. This person must acquire a thorough knowledge of the disparate information sources within the organization and the existing business intelligence infrastructure, and gain a fair understanding of the business processes along with the information requirements. Also, that analyst needs to be able to tap into a team of subject matter experts from business divisions, IT, and data analytics to complete a full picture.
Each KPI should be broken down into four elements:
1. Data source(s)
These elements together define the full scope and illuminate the different facets of a particular KPI. Information pertaining to each of these elements needs to be compiled in order to get a clear picture for each KPI.
Data sources would identify the high-level information regarding where to retrieve the information for a given KPI. Such high-level information includes database identification (specific data mart or data warehouse), Online Analytical Processing (OLAP) sources (specific cubes), data files (extract from legacy systems or data from external vendors), or existing reports and supporting sources (such as universes and objects). The image above illustrates a typical medley of data sources within a large organization that I refer to as the information biosphere. There are inherent challenges in integrating the myriad sources so that they seamlessly communicate with dashboard software to present all information in complete harmony.
During the process of identifying all of the KPI data sources within the information biosphere, loopholes that may exist within the organization's information delivery process may surface. For example, the process may reveal a lack of standards, lack of data validation, and data redundancies across various data sources. Therefore, it is a recurring phenomenon that the process of defining KPIs spawns a parallel effort into standardization, data mart development, data warehouse enhancement, and extraction, transformation, and loading (ETL) procedures. In other words, the process often initiates a full tune-up with overall checks and balances for maintaining a sustainable and balanced information biosphere within the organization.
Because the dashboard's function is to report on all aspects within the information biosphere, it is essential that the dashboard software have appropriate sensors to communicate with the different information sources. In the IT community, such sensors are referred to as database drivers, application programming interfaces (APIs), agents, adapters, and so on.
A comprehensive sensor checkup for the dashboard software would lead to a gap analysis between requirements and the capabilities of the software. In other words, if data sources include different relational databases, OLAP cubes, a vendor-specific reporting platform, and a vendor-specific Enterprise Resource Planning (ERP) system, it is imperative that the dashboard software be capable of communicating with all of these disparate sources. If the dashboard software falls short in connecting to a vendor-specific reporting platform, then an alternative is required to transform the information from its original platform into a more standard data representation that is readable by the dashboard software.
It is also important to note that a single KPI may involve multiple data sources. For example, determining the market share of a company's product may require the sales data from the enterprise data warehouse (EDW), and the total sale within that category as reported by an external vendor tracking the industry, reported through that vendor's proprietary reporting interface. If the vendor uses a proprietary homegrown reporting system, most likely any standard dashboard software would not be able to extract data from such reports. Therefore, dashboarding requires additional efforts to consolidate disparate data sources into a seamless data communication platform. This I refer to as the harmonization of the information biosphere.
About the Author
Shadan Malik is the author of Enterprise Dashboards: Design & Best Practices for IT published by John Wiley & Sons, portions of which appear here (reprinted with permission of John Wiley & Sons, Inc.).