I've just watched an excellent TED video from “Undercover Economist” Tim Harford.

In the video Harford tells the story of Archie Cochrane, a Scottish doctor, who while being held in a prisoner of war camp managed to figure out how to cure the prisoners of a mystery ailment (which he was suffering from himself) using a makeshift “split test”.

Later on after the war he challenged the medical establishment over the most effective post-operative treatment for heart patients. Again, using a simple test of the two different approaches to prove his case.

Harford uses the example (and others) to warn us against overconfidence in our ability to engineer solutions to complex problems and show that trial and error is most often the best route to a successful answer. You can watch the video here for more details.

The exact same thing is true in winning clients.

It's nice to think that someone out there can just hand us the answers to make our marketing work perfectly. But the truth is that what works in one business may well not work in another. Even if on the surface they look very similar.

So you need to use “best practices” as sources of inspiration to then test out, rather than taking them as the definitive answer to what your business needs.

I recently ran a test on my signup forms that resulted in a 59.3% improvement in optin rates. In other words for the same amount of traffic to my home page I'm now getting nearly 60% more optins.

And you can test things in the “real world” too, not just online.

Many people seem to be in a constant search for the perfect “elevator pitch” or introduction. They read one piece of advice and change their own introduction. Then they read another recommendation or “magic formula” and change it again.

But until they test it, they'll never know which format really works best or whether it makes any difference at all.

What they could do is try introducing themselves in one way at one set of events and another way at another. And keeping track of who got which introduction. Then measure how many follow-up calls or meetings they got from each introduction.

Not 100% scientific as there are many other variables. But do it enough times and keep good records and you'll begin to develop a decent picture of what works best for you.

Not all your tests will work, of course. I'm currently testing different text on the button on my website forms. The one I thought would perform best is currently getting 74.8% fewer optins than the control. But at least I now know not to switch to it, which I would have done if I'd just listened to others and gone based on my instinct.

What are you testing right now in your business. If you're not testing, you're not improving.