How to organise your application portfolio

How can IT managers tell the difference between projects and enhancements? And how can IT departments decide which upgrades will...

How can IT managers tell the difference between projects and enhancements? And how can IT departments decide which upgrades will offer the most benefit to users?

As global organisations drive towards stronger co-ordination between application development lifecycle tools and processes and portfolio management, IT suppliers are beginning to enable automated tool support.

With the advent of increasingly chaotic development environments, such as complex web services applications, co-ordination and customisation of enterprise resource management and datawarehouses and fewer resources, IT organisations are increasingly debilitated by function creep. As much as 40% to 60% of software enhancements for application development involves small projects of a short duration.

Cumulative costs result from the lack of organisational practices and prioritisation for establishing which enhancements are business imperatives and which can be bypassed. The expense incurred by diverting IT attention from business-critical enhancements toward creating functions of minimal or negligible benefit is exorbitant.

How do organisations determine the difference between an "enhancement" and a "project"? How do they retain a clear focus and establish a process for decision making and appropriate time frames for the execution of one change versus another?

During 2003/2004, Meta Group expects organisations to structure more effective processes and governance for prioritising and managing functional enhancement requests, driven by the need to cut costs and increase efficiency.

By 2004/2005, standard software configuration management products, combined with more mature governance and improved organisational practices, will enable users to co-ordinate change control and version management across applications.

By 2005/2006, stronger repository management will enable improved support for centralised control across the application lifecycle via centralised data stores and effective management capabilities.

A critical path for managing the plethora of user requests received by IT departments is establishing business priorities as the basis for decision-making and culling the enhancement wish list.

Although this area is driven by lines of business, both IT and business managers must buy into such criteria, which should be established in up-front, joint application development sessions, facilitated by a governance body such as an enterprise programme management office (where mature governance exists), or by appropriate advisory boards.

Decision-making criteria range from issues such as criticality of the enhancement to improving speed of customer support or product development.

Although major application releases may occur only once or twice a year, smaller requirements should be grouped into "point" releases on a regular schedule (every four to six weeks) that can be supported by quality assurance testing teams.

Incorporating a strong system to track requirements and assign priorities enables the development team to plan point releases with a small enough scope to be manageable by both development and testing groups. Requirements should be grouped iteratively by module.

As high-priority items are put into the point release, lower-priority items can be scheduled in a related module as appropriate (the low-hanging fruit concept). The wish list is thereby kept well pruned instead of growing into a long list of "neglected" items that can hurt IT staff credibility.

As IT departments establish end-user trust in their delivery of point releases at a quick pace (every four to six weeks), the number of "hot" or "must do now" requests will drop. This further increases the ability of the organisation to take a controlled approach (and exemplifies how an organic system's feedback loop helps to build stability).

From a process perspective, organisations should consider requests for production enhan-cements as a pipeline. From the pipeline, they should create portfolio releases of production versions of applications, meaning there will be "n" planned upgrades to the portfolio each year.

The pipeline is then set for a particular release. This approach assumes that the major upgrade is being treated as a separate project with clear scope definitions.

IT groups should use the prioritisation criteria established in the business/IT joint application development sessions to determine the business appropriateness of upgrading the portfolio with a piece of functionality, versus waiting for a new project to kick in.

The portfolio manager must have the leeway to refuse or accept which enhancements will be implemented. These are two separate (yet related) efforts which converge at the portfolio. Increasingly, we see a push toward co-ordinating automated project portfolio management with lifecycle tools (exemplified by Mercury Interactive's acquisition of Kintana in 2003).

Users should also evaluate and adopt standard software configuration management tools along with requirements analysis frameworks and testing. Automation of change management, as well as increasing co-ordination between software configuration management, requirements analysis, testing tools and application and project portfolio management, facilitates the ability of IT departments to efficiently prioritise and manage enhancements.

Meta Group expects to see stronger integration from suite players for such capabilities in conjunction with best-practice templates for functional prioritisation. Configuration management for ERP products is still lagging in the general market (niche players such as Quest and Mercury/Kintana provide support), with limited support available from the suppliers them- selves and general players.

Organisations must minimise code changes while upgrading to new versions of SAP, for instance, until the code base has stabilised. Within that timeframe, they should gather and prioritise code change requests, based on established criteria, to be poised for implementation once the new release has been rolled out.

Users should set standard criteria for prioritising application enhancement requests and establish automated tools to assist in the management of change requests because of increasingly chaotic development and customisation environments.

In 2004/2005, there will be increasing organisational maturity to facilitate governance through enterprise programme management offices and better co-ordination among software change and configuration management, testing, and portfolio management capabilities to facilitate increased efficiency.

Melinda-Carol Ballou is an analyst at Meta Group

Read more on IT project management