When someone writes a book called "Do It Wrong Quickly," it goes without saying that that person likes hyperbole. (And then I went ahead and said it anyway.) Of course, I don't mean you should be trying to do it wrong, but rather that you admit that most of what we try is wrong, and we need to measure the results so we can fix it. And I often tell people to change it, change it, and then change it again.

But there is a time to stop.

Here are three reasons to decide that a particular aspect of a campaign is OK as is, at least for now:

• You reach diminishing returns. In economics, it's well-known that you get most of the value of any action when it is first taken, eventually getting to the point that doing more and more returns less and less. When you are tweaking a campaign, and you see that your clickthrough rate soars with a change, but then further tweaking that change shows little improvement or even regression, then it might be time to work on something else.


• You have bigger fish to fry. Don't miss bigger opportunities just to perfect a small one. Even if you think that you can dramatically improve a campaign, it's not worth doing if you are missing the chance to improve another one that drives even more value to your business. Work on the most important things first.


• Excessive change is sometimes penalized. George Michie had an eye-opening post a couple of weeks back in the RKGblog called "PPC Copy Changes Can Hurt You"—he illustrated several reasons why constant churning of your paid search campaign can be counterproductive. In addition, you could imagine that daily updates to your LinkedIn profile or incessantly changes to your Web site architecture and design could cause feedback to be bad, regardless of the intrinsic value of any particular change in and of itself.

So, experimental marketing is a good thing, but only when it is measured in terms of business value. If your experiments are making big improvements in your most important campaigns, you're working on the right things. Like any technique, experimental marketing can be abused when it is approached as an end in itself instead of a way to improve business results. I write about this less because I suspect more companies are at the starting line deciding whether to run. What about your company?