Give me a better product, not more new features

John Carmack, the guy who programmed Doom and spent practically twenty years pushing the limits of realism in computer games, was asked in an interview “what new features the Doom engine would have.” He had a good answer: “More new features give you more product, but not a better product.” Speaking of greats of the computer age, Steve Jobs was an expert in this area. He was obsessed with it.

Update: the interview with Carmack

A user always wants more of everything — more features, more artwork, more levels. The assumption that hiring more people gives a better product is often incorrect. It might (but not always) get you MORE product, but not necessarily BETTER product.

As a cautionary example I'd point to what Atlassian has done, at least in Jira and Confluence. They do a million and one things, but none of them well… Anybody who's tried to write something in their editor will surely agree. To give the Australian company a break, I'm sure you can find plenty of products, physical or otherwise, that keep adding more and more features of ever less use just to set themselves apart from the competition.

I'll pick one more example out of many: Evernote. I was tempted for a moment to say Android, but that starts to sound like a holy war. Besides, Google didn't mess it up, much as I'd normally tear into Google. Samsung, HTC and whatever other Asian companies did that, with their fixed idea that people need two photo albums or calendars. Don't take those examples too literally. Substitute anything that turns up after you install a manufacturer's app bundle.

By the way, there's a direct relationship between the age of a product and the amount of unnecessary stuff added to it. Maybe the original definition of the product has changed ten times. Or maybe somebody started a “feature” and never finished it, so someone else tries building it again because they don't know about the first one. Maybe it's complicated and nobody wants to change it because they don't understand it. Then there are the customers. Yes, the one group that decides whether a product succeeds or fails. The rascals sometimes use features a little differently from how the engineers designed and built them, so we can't take them away. That puts us in a vicious cycle where one new feature follows the next. The product gets harder to understand and remember, less important and ultimately less valuable in the real sense of being something we'll pay for.

Doing product management well is hard work these days, because every damn person knows from common sense where each button or dialog should go, how it should behave, what it should say and above all what color it should be. If you're a developer and users or people relaying their wishes give you your specifications, you have no product. Zero, zilch, nothing. Unfortunately, you're well on your way to becoming exactly that thousandth product in line, hard to get a handle on and fighting for survival in a deadly cycle of new features and frustrated users.

Update: Medium CEO Ev Williams described it well in Medium is not a publishing tool

Understanding Blogger as we did at the time — as a software tool for creating and publishing web sites — we found ourselves in the race many software makers know well: Add features, get more users. Competitor adds more features, lose users. (Marketing and others factors have some effect, depending on the market. In blogging, that was minimal.)

Why am I writing about this? Everyone's obsessed with how few good developers are on the market. Compare that with how many people on the market are good at product management. With all due respect, if someone can't hire developers, I can't imagine what their product team looks like. I'll repeat myself gladly. Focus on the product. An interesting product will attract developers on its own.