I’m a huge fan of testing. If you’re running PPC campaigns and aren’t testing, then you’re not doing your job. Just because keywords in your campaigns are delivering traffic and conversions, it doesn’t mean your campaign is optimized. The entire point of optimization is to get the most conversions at the cheapest cost, and it’s impossible to determine what an optimal point is until you’ve been there, and then left it briefly. This is how testing works. The entire purpose is to take advantage of areas of opportunity, and sometimes the area of opportunity can come from an unexpected place:

Consider this: you have a campaign that hits its daily budget every day. You operate on strict budgets so you can’t raise it anymore, but in order to optimize for conversions you drop the bids on the non-converting terms.

It would make sense that with lower bids the number of impressions and clicks would decline, which would bring costs down, and maybe you lose a conversion or two from the lower positions.

A week later you go and look at the reports and notice that impressions, clicks, and even conversions went up! How could lowering bids mean more costs? The answer lays within the campaign budgets being set to rather limited daily amounts.

Some people have a difficult time grasping this idea because $100 in cheap clicks should be the same as $100 in expensive clicks. (Remember the old saying “a pound of bricks is the same as a pound of feathers”?). This is still true because the measurement–the buying power of $100–is accurate. And although the measurement might be the same, the value is not. After all, it takes a lot more feathers to reach a pound than it does bricks.

For example, if I had a budget of $100 a day with my cost per click bids set to $1.00, then I could afford 100 clicks. However, the actual CPC is rarely the same as the bidded CPC, due to quality scoring, placement, and other factors, so my 100 clicks might have only cost me $99.01. If this is the case, then I can’t afford one more click because at $1.00, the next click could put me over budget to $100.01. If over the period of 30 days I somehow always lose that last click and only spend 99.01, then my total expense for the 30 days will equal $2,970.30. The next month I decide to drop my bids from $1.00 to $0.40. The next day instead of my budgets shutting off at $99.01, I can still afford two more clicks, bringing my total spend per day to $99.81. Over the same 30 day period my total expense would be $2994.30, or a difference of $24.

If you take those 2 extra clicks a day and you have a 2% conversion rate, then you should have gained one additional conversion for $24. If your previous cost per conversion was greater than $24, then you came out ahead. If it was less than $24, then it would make more sense to raise your bids and give up the extra conversion gained.