Does it have to be perfect?

I just read a blog post by Ben Nadel stating “Don’t let great be the enemy of good”. At work, I am constantly struggling with “What can we do quickly and what would be perfect?” As a UX Designer, I know that even the slightest flaw can be a pimple right on the nose of your application. As a developer, I also know that “time-to-market” is important and that sometimes you need to release something that is good, even if it isn’t great.  So how do you know when it’s ready to release?

There are two categories of UX Design.  One is for low-frequency sites and the other is for medium-to-high frequency sites/applications.  An example of a low-frequency site is  You only go there when you want to buy this years Quicken or to get help with a problem using the program.  In this case, most visitors came once per year.  An example of medium-frequency site is Craigslist or linkedIn.  You visit these sites sporadically and they have more varied uses.  A high frequency site is often an application like Gmail or your company’s intranet.

In a high frequency site, the user is spending alot of time there.  They will get used to flaws.  They will find workarounds.  In this kind of application, you want to release as often as possible.  Anything “improved” is better than not.  The users will appreciate that the site is improving and will develop loyalty.

In a low frequency site, the user expects the experience to work.  They are not going to get used to flaws.  They will develop a long-lasting negative impression if things are bad.  In these case, you have to strive for something better.

So the question, Does it have to be perfect?   As a UX Designer, the answer is “Only if you want to have a product people LOVE using.”  As a developer, the answer is, “No, we will have to progressively improve it.”  The problem with the latter statement is this.  If you promise to progressively improvement something, you have to remember to go back and do that.  The thrill of new features is always stronger than the promise of improving past features.

As an industry, we have to do a better job of going back and improving the old stuff.

Whatya think?