More than anything else, starting a small business has been a learning process. When I learn valuable lessons, I try to pass them on to others who may be in the same boat.

I recently tried Yahoo banner ads—not the Adwords, but rather the graphic ads. They have some beautiful templates, and working with the product itself was pleasant and fun. Then, when I was ready to go, I was presented with a choice: Pay 1.30 per 1000 impressions (CPM) or pay 90 a click (CPC). So, essentially, you pick one variable to pay for, and let the other variable float.

My first opinion was, try per click. Yahoo argued against CPC and for CPM, saying something about the different networks that will carry either one. I figured, perhaps naively, that they were probably giving me good advice... they want me as a repeat customer, so they want to see my campaign be successful?  I decided to test each, and set up a nice head-to-head race.

The results were conclusive and painful. My cost per click was obviously fixed for the CPC campaign; for the CPM campaign, I was paying a fixed amount per impression and a floating amount per click. This cost per click floated alright... up into the stratosphere, in excess of 700% the cost per click of the CPC campaign. The cost was simply staggering. If you try something like this, please keep my experience in mind. If you insist on trying CPM, something I will certainly never do again, consider running a tiny test of both methods in a head-to-head comparison. Interestingly, my click-through rate (the number of times my ad was clicked, per 1000 impressions) was THREE TIMES higher for the CPC campaign than for the CPM campaign. Identical advertisements, btw.  So, they apparently know where to go to get the clicks, when that’s what they’re making their money from.

The benefit from this campaign was simply this: I now know to not try it again. My curiosity is sated. A significant amount of money later, Yahoo reported about 250 clicks, Google Analytics reported only 60. I’m not sure who to trust, though I suspect that Google Analytics may not report visits that are shorter than 3 seconds. (Anyone know? Please let me know, if so.) I hate to say, “click fraud”, but I do wonder.

In Yahoo’s defense, they would tell you that only a large campaign works well. They would say that repetition is the key… that Geico’s non-stop advertising makes sense, that a magazine publisher needs to insert not just 1 or 2 subscription cards into a magazine, but 5 or 6 or more. (I personally always found that amusing. "You didn't leave an impression on me the first 6 times a subscription card fell out on my lap, but wow, this 7th card is just really something special!") Perhaps this is all true.

But, where do we go, when we don't have the budget to fund a huge, repetitive campaign? So far, and I'll let people know if I learn anything important here, but Facebook is working pretty well. Based on my Yahoo learning experience, I'm paying per click and letting my cost per impression float. So far, it has sunk to about 20 cents per 1000 impressions. And you can tell the difference with these ads. I actually clicked on one, an intriguing little oddity called strawbalegardens.com. The site is obviously a labor of love: someone with a cool idea, trying to sell a book about it and apparently nothing more, who obviously just enjoys what he's doing. He's using Facebook, with a budget of possibly $5 a day, to get the word out. Again, I'll update regarding my Facebook experience when I have know something conclusive, but so far, so good.