PerformancePoint Server 2007 is both a server and a platform, offering three different areas of BI functionality: monitoring, analysis, and planning. The monitoring function lets users view the health of the business or of a particular department quickly and easily. The analysis function provides tools to explore data in depth. The planning function allows organizations to easily create strategic plans, forecasts, and budgets, which can then be fed back into the monitoring piece for comparing actual performance against those forecasts and plans. Part of PerformancePoint Server is a server, running a background process that handles workflow for the planning function. It’s also a platform, providing Web services for easy programmability and customizations.
Microsoft recently launched a new product as part of its business intelligence (BI) arsenal: Microsoft Office PerformancePoint Server 2007. PerformancePoint Server is interesting because it contains three distinct pillars of functionality in a somewhat integrated solution that supports a variety of users. I say “somewhat integrated” because the pieces aren’t as tightly integrated as originally envisioned, but this doesn’t detract from the power and usefulness of this first version of PerformancePoint Server. We should expect much tighter integration in future releases.
So just what is PerformancePoint Server? It’s both a server and a platform, offering three different areas of BI functionality: monitoring, analysis, and planning, two of which are integrated into one installation and another that requires a separate installation. The monitoring piece can access data stored in several different formats, but the most common data source is SQL Server Analysis Services. The data in Analysis Services comes from a variety of sources in the enterprise, including financial systems, customer relationship management (CRM), ERP, HR, and other systems. The analysis piece accesses data in Analysis Services cubes, and metadata for both the analysis and monitoring pieces is stored in SQL Server 2005 databases and displayed primarily in SharePoint.
The third piece, the planning server, stores data in SQL Server 2005 and generates Analysis Services cubes. It’s used to provide budgeting and forecasting for the organization. Data used to start the planning process most often comes from the current accounting and financial data stores, although Microsoft Excel is another popular data source.
PerformancePoint Server Details
Because Microsoft has created just a single version of PerformancePoint Server, you don’t need to distinguish between standard or enterprise editions at the moment, although whether this will change in the future remains to be seen. The retail cost for PerformancePoint Server is $20,000 with a client access license of $195 for each user. An optional external connector to allow non-employees to connect to PerformancePoint Server is $30,000. This connector is a license that lets people from outside the organization connect to a SharePoint site that the organization makes publicly available; the license avoids the need to have a CAL for users that are customers or vendors.
PerformancePoint Server relies on several other tools from Microsoft. Given that it’s part of the BI product stack, it’s no surprise that SQL Server 2005 stores PerformancePoint Server’s data. Analysis Services isn’t technically required for the monitoring and analysis pieces, but many businesses use Analysis Services cubes as the data source for their monitoring and analysis. The planning module does require Analysis Services to build cubes of the models. SharePoint is also required—either Windows SharePoint Server (WSS) or Microsoft Office SharePoint Server 2007 (MOSS), as there’s no difference in the functionality of PerformancePoint server on WSS or MOSS. For the monitoring and analysis pieces, Microsoft IIS is required for previewing results. Additional required components include the Microsoft .NET Framework 2.0 and ASP.NET 2.0 AJAX Extensions 1.0.
From a hardware standpoint, the requirements for PerformancePoint Server are lightweight. The planning module runs on one single-core processor with 1GB RAM, and the monitoring and analysis modules combined will run on the same system. A more realistic setup is a two-processor dual-core machine with 2GB RAM for the planning server or for a combination of the monitoring and analysis modules.
Now that you’ve learned about the pieces of PerformancePoint Server and what it needs to run, what exactly does PerformancePoint Server deliver? Let’s look at the three pieces in depth.
The monitoring functionality of PerformancePoint Server aims to help businesses easily visualize their performance through highly intuitive, graphical tools. Key Performance Indicators (KPIs) are displayed on scorecards and dashboards, which make up the bulk of the monitoring features in PerformancePoint Server. The monitoring piece of PerformancePoint Server is relatively complete, as it’s really the third version of Microsoft’s scorecarding product, replacing Microsoft Office Business Scorecard Manager 2005. Recall that the data for scorecards comes from existing systems such as manufacturing, financial, human resources, and other source systems in the organization. Scorecards rely on data warehousing, part of which involves consolidating disparate data from across the enterprise.
KPIs are numeric values that measure performance and indicate how that performance compares to the expected value. KPIs contain several different items, the first of which is the actual value. Second, KPIs contain a target value, which can be considered the plan, budget, or forecast value. It’s what the metric should be for the time period being viewed. KPIs also contain an indicator of the strength or health of the KPI, using symbols such as stoplights (with green, yellow, and red indicators), smiley faces, gauges, and so forth. Finally, a KPI might include a trend, which shows the direction of the actual value.
The KPIs measured vary by company. Some KPIs are financial, such as sales, profit margin, cost of goods sold, and hours billed. Other KPIs might include items such as customer satisfaction ratings, market share, employee turnover, employee training hours, inventory turns, defect rates, and more. Each company will track different items it considers strategic to its health.
KPIs by themselves are somewhat useful, but they’re even more useful when placed on scorecards. A scorecard acts much like a student’s report card, which doesn’t show the score of individual assignments but instead shows an overview by displaying a single grade for each class. Similarly, a scorecard shows a summary of many individual numbers by showing KPIs for broad topics. The advantage is that users, often executives and business decision makers, can get a feel for the health of the business in a single glance.
Scoring makes it easy to rank different divisions, products, and other elements used to monitor performance. Imagine that one company is keeping track of its customer satisfaction rating. If the company has a target satisfaction rating of 95 percent, and one division has a rating of 82 percent while another has a rating of 98 percent, the division with the higher rating will clearly have a higher score. Exactly what the scores indicate will depend on how the business chooses to calculate customer satisfaction (or whatever the score represents), but clearly those divisions with better performance will have higher scores. KPIs can roll up into higher KPIs and can be weighted so they contribute the appropriate weight to higher-level values. For example, there might be a high-level KPI named “Customer Value” that includes detailed KPIs for Customer Satisfaction, Customer Returns, and Market Share. Each of these lower-level KPIs would have a score that could roll up into an overall Customer Value score.
PerformancePoint Server provides a graphical tool, the Dashboard Designer, for creating KPIs, scorecards, reports, and dashboards. A dashboard is merely a container that can hold many scorecards and reports. The Dashboard Designer blends the monitoring piece (scorecards) with the analysis piece (reports). Dashboards are deployed to Share- Point for viewing by end users. Figure 1 shows an example of a dashboard with a scorecard and report.
The analysis functionality of PerformancePoint Server is sometimes unofficially referred to as “analysis lite” because the amount of true analytics that made it into the first release is limited compared to Microsoft’s original plan to incorporate ProClarity products, which it acquired in 2006. At the time of the acquisition, ProClarity offered the leading third-party analytics application for accessing Analysis Services cubes.
The analysis functionality that did make it into PerformancePoint Server centers around reports that can be created in the Dashboard Designer. The two primary reports are the Analytic Grid and the Analytic Chart. Both are thin-client reports that let users view data and click to drill down to lower levels of detail. For grids, there are standard plus and minus signs to expand or collapse regions, while the chart allows users to click bars or pie slices to see additional data. These reports contain limited analytics features but meet the needs of most users. More powerful and flexible reports exist, but some of these reports rely on the Office Web Components, which require both a client installation and a client license.
To fill the current gap in the analysis side of the product, a license of PerformancePoint Server will include a license of ProClarity. ProClarity has a Windows client version and a Web server version, both with nearly identical capabilities that let analysts and power users examine, pivot, drill into, and display data. The tools are designed to support all manner of data analysis, and users can add and remove elements as necessary to work with the data. PerformancePoint Server offers a report type that shows a ProClarity report, called a view, and ProClarity views can be added to a dashboard, which Figure 2 shows.
The planning module is designed to support the frequently onerous process of planning, budgeting, and forecasting. The process starts when models are created that specify how the business or its different divisions need to create their budgets or forecasts. Users then can create budgets or forecasts for their area by using a familiar tool, Excel. These plans are then rolled up to the corporate level, allowing more people to create more detailed plans and still conform to the overall business strategy. In Figure 3 you can see an example of how users interact with the planning module by using Excel.
Budgets and plans can go through the organization’s specified workflow process via a service within the planning module that runs on a central server. Task assignments, submissions, notifications, and approvals are handled through email, SharePoint, and Excel. From a security standpoint, users have access only to the data for which they have permissions. All submissions and changes are tracked in audit logs to comply with stringent financial regulations.
One of the features of the planning module is that the plan data is stored in Analysis Services cubes. This data can then be used as the target value in the KPIs in the monitoring module of PerformancePoint Server. The planning piece feeds back into the business monitoring module, helping tie together the vision of a single tool that supports the entire process of monitoring, analyzing, and planning a business. While the data can flow back into the monitoring module as the target for KPIs, and most users of the planning module will use Excel, the actual models are created with a separate tool, called the Planning Business Modeler. Whether the Planning Business Modeler and the Dashboard Designer will ever be merged remains to be seen.
A Tool with Value
PerformancePoint Server can require multiple installations and uses different tools depending on what is being created. Despite the imperfect integration of the parts in this first version, the overall vision is sound and the tool delivers tremendous business value.