I have a confession to make.

I've been a judge at Cannes. My work has been in Communication Arts and the One Show. But until I was 34 years old, I'd never produced a decent TV commercial.

The fact is I never had a formal ad education. Which means I primarily learned the hard way - through making plenty of mistakes.

I always thought a good ad was one the client liked and approved of. But it slowly dawned on me that no one ever talked much about my ads at family gatherings - or anywhere else. Not even my Mom seemed to take much interest.

That's when I finally figured out that ads can't be good unless they get noticed in the first place. And ads that get noticed tend to avoid certain pitfalls.

Mistake Number One: Cramming too many messages in your spot. This is incredibly obvious, but it still took me a long time to realize that the best ads say one thing well - and only one thing. The more messages, the more likely your ad will be ignored.

To help our clients understand this simple fact, we came up with an analogy we call the Ping-Pong Ball Theory of Advertising. If I were to throw one ping-pong ball at you, you'd probably catch it. If I were to throw five at you at the same time, you'd probably catch none of them. Most ads have five ping-pong balls. Which is why so few messages stick.

Mistake Number Two: Presenting multiple-choice creative. Clients are only human - if you give them a safe choice, they'll probably take it. Our rule is simple: We'll show clients any ad we'd be happy to produce. During the development process, we show rough ideas one at a time to everyone in the office - including the mailman, the couriers, and people off the street. We ask them if they understand the ad, find it relevant, and think it's original. Any ad that gets three "no" votes to any of those questions gets scrapped. Usually only one campaign ever survives this grueling peer review gauntlet.

Mistake Number three: Using research at the wrong time in the process. We're big believers in using research early to help understand consumers and define strategies. But we believe testing ideas at the concept stage is a guaranteed way to ensure mediocrity. Focus groups tend to reward the tried and true - not ads that are surprising and different. Without actors and music and nuance, it's impossible to separate the mundane from the magical.

In over five years at Rethink, no ad has ever been pre-tested. How can we get away with this? Partly because every concept the client sees has made it through our peer review process. They don't need to worry about basic comprehension, because we've made sure every ad passes that test.

Mistake Number Four: Presenting tight storyboards or animatics. During initial presentations, clients should be judging ideas - not executions. The best ads are like stories - and you shouldn't need a storyboard or animatic to tell a good story. Let the client picture the scene in their own mind - instead of getting hung up on details. At this point, camera angles and blocking should be the last thing you're worrying about. Which brings me to: