What Matters for Application Portfolio Management?
To be most effective, Application Portfolio Management depends on current, accurate, and rich insights into IT and your business. As discussed in ‘How to measure Application Portfolio Management’, data can come from analyzing source code, stakeholder surveys, and other technologies. As a result, APM lends itself to automation via software tools. In this section, we’ll look at the characteristics that you should require in an Application Portfolio Management solution.
Collect Technical Metrics from the Application Portfolio
An important baseline for managers must be measures of their application portfolio’s size and complexity. Metrics like function points, cyclomatic complexity, essential complexity, and others provide KPIs for tracking the inflexibility of your application portfolio. This should be standard in your Application Portfolio Management tool.
Because your application portfolio likely encompasses a broad range of environments – from COBOL to Java – the tool should have the ability to assess these diverse languages. Further, because standards will be different from company to company, you must be able to customize the collected metrics to conform to internal policies.
Survey Stakeholders to Understand Value and Risk
Understanding business priorities and opinions of stakeholders is critical. If you have two equally complex applications that are draining similar amounts of resources, it should be clear that you would direct resources to the one that generates more value for the business. But it is not a straightforward determination, and is more accurate when decision makers are asked to quantify the value and risk of processes, platforms, geographies, and other groupings.
To effectively survey stakeholders, users can rely on spreadsheets or paper documentation. However, it is clear that a more effective approach is to survey users via a browser-based interface. This allows users to quickly construct and distribute questions to a geographically distributed team. Responses can be immediately tabulated, keeping data as current as possible.
Collect Measurements from Other SDLC Tools
Gathering measurements from other technologies has advantages and drawbacks. The advantages are immediately clear. It provides a more complete picture of the application portfolio, and richer data for making better management decisions. The drawback, however, is that it can be difficult to gather data from such a range of tools.
Again, automation plays a role in simplifying data collection. A manual solution would likely yield out-of-date data and would consume too many resources in the process. Instead, collecting data in an automated fashion is more effective. This necessitates that your solution has an open repository for storing the information and interfacing with other standards-based SDLC tools.
Combine Metrics into Business Metrics
Metrics are useful because of the business decisions that they support. As a result, you must be able to take metrics like cost, complexity, value, and risk and combine them into KPIs that help guide your investment decisions. This means that you should be able to manipulate data, combine metrics, and perform other math to generate useful measurements.
As discussed in the “How to measure Application Portfolio Management” section, you may wish to combine the quality of a development team’s output with information about the frequency of change of the code that they manage. This could help determine which projects require additional resources to be allocated.
Distribute Application Portfolio Management Data to Right Users
Application Portfolio Management data is only useful if it is presented to the right user at the right time. A CIO requires portfolio-wide information abstracted to a useful level to identify priorities. A development team manager will require a more focused view on just her team. A business analyst will want information about the applications that automate his business processes. And an outsourcing manager will want information about the applications controlled by his service partners.
Access to this data could come from browser-based dashboards that are customized to the information needs of the user. They should allow a user to ‘drill-down’ and focus attention on areas of interest. But Application Portfolio Management is more than that. It should also be thought of as a ‘feed’ of information to the right users in the SDLC. So, as a developer needs information about whether his changes affect the architectural quality trend, APM should provide that targeted data. This sophisticated APM can have huge productivity benefits, but does require a flexible solution with an open repository.