There is a school of thought in the SEM world that believes the cure for all evils is simple: Do more! More keywords, more ad text, more landing pages, more targeting! Sometimes doing “more” is necessary and valuable; often, however, the result is “diworsification,” where “a business that diversifies too widely, risks destroying their original business, because management time, energy and resources are diverted from the original investment.”
Consider, for example, a company that sells one product -- let’s just call it a “blue widget.” Odds are high there’s somewhere between 15 to 20 keywords that drive 90%-95% of the non-brand search volume for this product. But some energetic SEM manager decides that buying 10,000 long-tail keywords like “best low price blue widget for sale online” is going to drive hundreds of low CPA conversions. Instead, the account spirals out of control as Google bounces queries from one keyword to another, and odd broad matches cost the account oodles of money. That’s diworsification.
Or the newly graduated English major, hell-bent on finding the absolutely perfect ad text, who decides that each ad group needs at least 10 ads tested at once. The ad group ends up struggling as too many clicks go to ads that just aren’t working, and the time required to figure out which ad is the right one is extended into weeks or months. That’s diworsification, too.
Of course, the opposite of doing too much is doing too little, or doing nothing. I recently audited a $250K/month AdWords account and noticed in the change history that the agency had spent a total of eleven minutes on the account over the last 45 days! Standing still is the equivalent of moving backwards in SEM. And not doing any testing is a recipe for disaster.
It’s also important to always be testing a little to make sure you don’t miss the next big thing in online marketing. If all you are doing is desktop SEM these days -- and you aren’t testing out mobile campaigns, or Facebook, or display -- you need some diversification, and pronto!
As you probably expect, the answer is somewhere in between the two poles. Too much testing done without a plan -- and done too quickly -- drives unnecessary complexity and increased opportunities to make costly mistakes; too little testing results in stagnation and an inevitable decline in performance.
Every account doesn’t need 250 million keywords, nor does an account deserve to lay fallow for days on end. Find the right balance between tending to what already works and experimenting with what might work better, and you will reach SEM nirvana!