the feature creep conundrum

James Surowiecki (of Wisdom of the Crowds fame) has a great article about Feature Creep in this coming Monday's May 28 issue of the New Yorker. In the article, he adroitly addresses the painful situation that technology producers have found themselves in now that design cycles are relatively short, innovation is relatively cheap and the consumer feedback loop is so short:

This spiral of complexity, often called “feature creep,” costs consumers time, but it also costs businesses money. Product returns in the U.S. cost a hundred billion dollars a year, and a recent study by Elke den Ouden, of Philips Electronics, found that at least half of returned products have nothing wrong with them. Consumers just couldn’t figure out how to use them.

Surowiecki points out that a lot of Feature Creep problems stem from the internal-audience problem where the developers themselves don't use the very products they're developing and therefore have no idea how crappy the end product turns out.

I learned at AOL to combat the internal-audience problem by making the producers use the actual goods they're making (aka "eating your own dog food"). Back in 1998, when my team created the 1-2-3 Publish and Easy Designer home page publishing tools, you bet I used them to make and maintain my own AOL Hometown page (still there after all these years. Ah, good times).

The number of features we didn't include in the basic publishing tool was huge, and not all of them fell off the list because of scarce development resources... we had constant battles to strip down the 1-2-3 Publish tool to be as lean as possible so newbie web publishers wouldn't get distracted by gizmos on their way to creating their first web page.

Of course, as a product manager, I cheated by throwing all the gee-whiz features intended for 1-2-3 Publish instead into the specs for Easy Designer, the WYSIWYG publishing tool. The price we paid for including all these things was comically bloated html code in the resulting published pages. As long as you don't view source, the products not all that bad.

I still make sure to use the tools I produce, but my primary interaction with feature creep now is as a consumer. I fancy myself to be a geek and take pride in making as much use of features on my gizmos as possible, only if I have time to study and learn and ham-fist my way through menus of options.

I've come to expect that part of the purchase equation I've internalized is that I'm willing to pay for high-end stuff and yet use only a tenth of the features available on that stuff. Case in point, my Harman Kardon AVR245. Of the 70 buttons on the remote, I use only 5 of them (power, volume up, volume down, FM, HDMI) when I'm not just manually pushing buttons on the receiver itself. That fact in itself tells you just how much of the capacity of the AVR245 I'm using. However, I love the sound quality of the AVR a lot, and even that's an understatement.

I'd like to say I bought the AVR245 specifically because of its sound quality and not because of all the cool kick-ass features I'll some day use, but that's not true. I bought it for the feature set (and the sound quality!), and that's exactly the conundrum Surowiecki talks about in his article:

as numerous studies have shown, people are not, in general, good at predicting what will make them happy in the future. As a result, we will pay more for more features because we systematically overestimate how often we’ll use them. We also overestimate our ability to figure out how a complicated product works.

Maybe I'll spend Memorial Day learning how to use another button on the remote.