Sunday, January 13, 2008

Google AdWords 101: Landing Pages

Launching a PPC (define) campaign is still the best way to advertise online. You can measure your results. You can make changes on the fly. You can control your destiny.

Last time, we looked at the basic steps to set up a Google AdWords campaign. Now let's look at how to test your results with an easy landing page test.

A/B Landing Page Testing

With your paid listings in place, it's time to use A/B testing to perfect them.

In Google, you can have multiple ads in any ad group. When you're ready to do an in-depth keyword analysis, isolate the keywords with the highest CTRs (define) and the largest impression numbers in their own ad groups. This allows you to write two creatives for a specific keyword and test the results. The code is similar to what we discussed last time, www.yourdomain.com/?engine=AdWords&keyword=blue+widget, with addition of "+1" or "+2" (numbering the different creatives) at the end of the link: www.yourdomain.com/?engine=AdWords&keyword=blue+widget+1. This way you learn which ad for this keyword sends the best converting traffic.

The process actually is a little more complicated if you want to do it efficiently. You now have two major numbers to work into a further set of calculations. The first is the CTRs for each term. I usually turn off optimization during the first round of testing, so each ad gets an equal number of clicks.

You then calculate the CPA (define) for each ad. Divide total CPC (define) for the ad by the sales it brings in.

Sometimes the CPA differences jump out at you, making it tempting to let the winner be your chosen ad before looking any further. This could be a big mistake. It's one of the parts many people overlook in the battle to perfect paid listings. When you have a significant difference in the CTR and CPA, you must explore what each can mean to the cost of doing business.

Let's take an example. Blue widget's creative 1 generates 80 clicks from 1,000 impressions at a cost of $0.50. The CPC can vary a little, but for easy understanding and calculating, let's say both creatives have the same CPC and a CTR of 8 percent. Creative 2 generates 40 clicks at the same $0.50 with a 4 percent CTR.

The tracking shows us that creative 1 gets eight sales (conversion to impression rate: 0.8 percent) and creative 2 gets five sales (conversion to impression rate: 0.5 percent). Thus, the CPA for blue widget creative 1 is $5: $40 CPC divided by eight sales. The CPA for blue widget creative 2 is $4 : $20 CPC divided by five sales.

It's true that creative 2 costs less to get a sale. But since creative 1 generates more sales from the available traffic, the question of quantity versus quality must be taken into consideration. As long as both are profitable, you need to do more calculations.

If the gross profit per blue widget sold is $8, the profit per sale is $3 for creative 1 and $4 for creative 2.

But what about the number of impressions? Given the limited number of searches for any keyword, the CTR and profit per sale can be critical.

Let's say there are 50,000 searchers. Creative 1 generates 400 sales at the conversion to impression rate of 0.8 percent. Creative 2 would generates 250 sales from the same 50,000 impressions. Creative 1, then, generates 400 sales at $3 profit per sale for $1,200 for every 50,000 impressions. Creative 2, meanwhile, makes $1,000 from the same amount of traffic. (It gets a little easier the more you do this.)

The next step is to test improving the CTR and the CPA. With advanced Web analytics, you can then do visitor path analysis to see where visitors go on the site as a group when they come from a certain ad set. This allows you to make further changes to links on your page and where you land searchers when they do click.

Any successful PPC campaign has to work with the site it's sending traffic to. The more the two work in conjunction, the better the results will ultimately be.


No comments: