“You don’t want to run duplicate keywords, because you’ll end up competing with yourself”
I’ve heard variations of that statement dozens of times (including from some colleagues at Google back in the day) but I believe it’s misleading. Here’s why.
The argument usually goes something like this: Having multiple keywords vying with each other only jacks up your own prices unnecessarily, and effectively adds to your own competition before you even get involved in the real auction.
Most of this is true, but none of it is a problem… precisely because this internal competition is happening before you get involved in the real auction.
This internal competition is based on the same ranking factors that you take with you into the ad rank contest out in the wild. Whichever of your keywords boasts the higher ad rank at the time of auction, is put forward to represent you (think of the prelims for the Eurovision Song Contest…. OK maybe don’t).
That internal ‘contest’ doesn’t cause any price inflation (except for the fact that your keyword with the higher bid may well win out). The end effect on actual CPC will be determined by the real ad rank competition to come.
There are reasons not to run multiple instances of the same keyword. It fragments data, affecting both your ability to assess and optimise, and your keywords’ ability to build quality score.
But ‘competing with your self’ and raising your CPCs is not one of those reasons. The case against duplicate keywords is overblown in that respect.
The internal ‘competition’ part of it is actually quite healthy from your account’s point of view, as it helps make sure your ‘fittest’ keywords go out into the wilds to bring home the bacon.
You can use this to your advantage… though once you’ve let them compete – and established which keyword instance fares better – then yes, it’s usually best to leave that winning keyword alone (to see if it can impress the voting public!).