Paper: “Escaping the design traps of creeping featurism: Introducing a fitness management strategy”

Authors: Dong-Seok Lee et al., Cognitive Systems Engineering Lab, Ohio State U.

[MC: I've summarized here some useful bits of a longer peer-reviewed paper included in the UPA conference.]

Feature creep is a process of creating complexity over a series of design decisions. How can we be aware of when feature over-complexity starts to dominate, and how do we head it off? Their suggestions:

• Focus internally on product evolution: When companies call in usability consultants, they focus on a one-time fix. Usability inside companies must have wider points of view in order to manage the evolution process of their products rather than trapped in short-term schedule pressures.
• Check how users cope with complexity: Not many studies investigate how users deal the complexity of UIs. Rather than just measuring task performance data, seek to understand what difficulty users confront while interacting with products and what strategies they deploy.
• Ask what will be made worse by adding new features, not just what will be made better: Evaluation usually focuses on the bright side of changes, but there may be a dark side.
• Look for what to remove as well as what to add: In many product development organizations, everyone looks for new features to add, but no one considers what to remove. It is essential that development teams check what features are not used or are used infrequently.
• Dig for causes of new or worsening usability problems: these could be symptoms of feature creep.
• Focus on the value of products (solutions) rather than features: What is important is the overall value of products for users, not their new features. Experimental settings or survey sheets are often designed to show the merits of new features.
• Analyze user complaints from web sites, sales people, or service calls: When complexity dominates, users look for simpler products or alternatives.
• Change-manage the UI and the reasons for the drift: You must know the details of design changes, intent, and impact in order to detect and respond to complexity.

[Note: This rings true from moaning I'm reading on another list, about how SQL Server 2005 has lotsa nice new features for big customers, but at the expense of goobering up the UX for existing ones.]