Feature creep is a symptom. The cause is lack of a coherent focus for the product development effort, no design target. Everybody, even casual or nonusers can suggest features, as one feature is just like any other (the only variable is code considerations, not suitability to task or user goals).

This happens when there is no mechanism in place to say "no" to a feature request. Coupled with an inability to view feature sets with design integrity, and the project is always following around users instead of leading them. Quite a few companies will not add a feature unless the user asks for it. Since no user asks for coherent design, the company can't get behind this concept. This is not always a coder problem, but can just as easily come from any part of the company.

Posted by: on June 23, 2003 10:44 PM

---

 

Unfortuntately, people buy features and the more features, the better value. I think Nokia has slipped into features land. What do other people think?

Posted by: daniel szuc on June 24, 2003 07:42 AM

---

 

I think there's a lot of buck-passing going on whenever the subject of overcomplication and feature creep comes up. The coders always point to management and marketing, saying "We didn't want to add 100 new features, but they insisted to make it more marketable!" And the bosses and marketers shrug and say, "That's what customers want. They're not going to pay for a new release that just improves the interaction. Customers demand more features, so that's what we have to supply."

They're both conveniently ignoring the fact that this is a leftover tradition from an older industry with a different customer base. The early days of software focused squarely on hardcore programmers as customers, and those people DID want more features and more control in every upgrade. So "new release" translated to "new features" as a standard practice.

The industry has changed now--coders and tech geeks aren't the primary customer base any more--but inertia is still driving software developers to do the same thing... except it isn't working any more. Customers are disgusted with overcomplicated, bloated software that's hard to figure out and has hundreds of options they don't want or need. Many people now buy upgrades only when they know tech support on the old version will expire, simply to avoid the hassle of wrestling with another layer of new and modified features they never asked for.

Posted by: Calybos on June 24, 2003 08:02 AM

---

 

Excellent points, Calybos.

The impression is a feature, any feature. There are no good or bad features in feature land. Coherency in interaction design is too abstract a concept, not for users, but for development. (David Pogue has a nice article on cell phones where the criterion for evaluation was simplicity.) Again, evaluators don't lead the cause by explaining the value of simplicity, or even design integrity.

Most apps, products and so on are pieces badly joined. You don't see, for example, a testing lab made up of office workers with a deadline to meet. Nobody tests apps by putting together a real project, or timing the most basic user task. It has always been about how fast, or obnoxiously distracting the report or presentation it can make. Not why anyone would want to use the technology to accomplish anything important.

There is blame enough to go around. But one piece of the puzzle is the idea that all techology starts off as a toy, then becomes an important tool. They have the first part down perfectly. They haven't figured out the technology doesn't stay a toy forever.

Posted by: on June 24, 2003 08:33 AM

---