Using A/B Testing to Develop Powerful, Magnetic Ads

You will increase your conversion rate, clickthrough and ROI using A/B Testing

By WG Moore
Web Stats Gold

My goal in this article is to show you how to use A/B testing and your website traffic statistics to develop powerful ads that grab your visitor's attention and draw them into your web site like a magnet! Done right, it will weed out the researchers and those looking for freebies and bring in visitors who are genuinely interested in your products and services. This means that your pay per click ads will enjoy a higher click-through rate of people ready to buy. It also means your ROI and conversion ratios will skyrocket.

So, What is A/B Testing?

We will have to make a few assumptions here: One is that you have a web statistics service that is capable of measuring visitor traffic. This is an absolute necessity in running an e-commerce business. Preferably, the service will have Ad and Sales Tracking statistics built in. Be sure and ask. You can get by with basic log analyzers and free web stats services, but it is a bit slow and somewhat difficult.

The other assumption is that you know how Pay Per Click works. Most of the SEs like Google and Yahoo provide a tracking code to paste into the landing page so that you can count clickthroughs. The click through rate is how many people actually saw your ad and clicked on it. It is a simple numeric count. For the purposes of simple A/B Testing, this is enough.

Let's Get Started.

First do a little research. Go to Google, Yahoo or any other search engine where you plan on running ads. Enter your most powerful keyword or keyphrase and look at the results. Since we are concentrating on ads for the moment, take a look at the ppc ads that appear on the right, or maybe the top. Which ad drew your eye first? Was it the top-most? Or was it the most intriguing? Now, which ad do you want to click on the most? Make notes of these and also write down what you found most attractive about them.

I am willing to bet that the ad you like the most is one that stressed some benefit rather than some feature of the product, right? The fastest way to get someone's attention is to show them how they will gain from your product. Always stress Benefits, Benefits, and more Benefits!

You are paying for each click on your ad, so make it clear what you sell and how it will benefit the visitor. But don't just show the name of your company and list products. That is a sure way to go broke. Use words that clearly show how your visitor will benefit by following your link.

In this article, we will concentrate mostly on the testing procedures. So for now, we will assume that you have your killer headlines ready. Yes, I said 'headlines', plural. Never write only one. You will need many versions using different words.

What we think is the perfect headline is probably not what works in print. Others do not see through your eyes, so make sure you tell it to them as clearly as you can. And don't hesitate to get inspiration from other ads you see while you are doing your research.

Boil it down to the minimum number of words that still build enthusiasm and show benefits to your visitors. Also, be sure to use key words and phrases that are found in the body of your landing page.

So by now you have written the perfect Pay Per Click Ad. Make a copy. You should now have ad 'A' and ad 'B', both precisely the same. Now change the headline of ad 'B', using your second choice most powerful headline.

Alright, time to set a budget. Let's say that you can afford only 100.00 for each ad. If you want to try to use less, and you can get a top placement position, by all means, do so. Remember, you don't have to let the ad run for long. The Internet is a fast paced medium.

Now submit both of your ads the same place. Since this is A/B testing, we only want one variable between the ads. Submitting to two places might seem like a better idea, since you would get more exposure, but then you would have two variables, and that is not our goal here. Our goal is to test until we find the best possible ad. Then we can go for exposure.

Also, you don't have to place your ads in the most expensive, high traffic location, just some place with reasonable traffic and a good page one location. If you get no response, then move both ads to another location. You have just tested one location. If the second location also doesn't draw traffic, dump one of the ads and try again.

Once you have a clear indication of which ad draws the best, change the headline of the worst performer. Run the test again. Do not change anything else in the ad. We only want to test one thing at a time. When you are happy with the results of the headline performance, make changes to the body of the ad and again repeat the test.

Remember, you must test, analyze, change and test again, and again and again. As you refine your headlines and text, your ads will begin to draw more and more qualified BUYERS to your site. You will finally reach a point that further changing and testing will only lower your performance. Let it rest for a while until you see your clickthrough rate begin to drop. Then it is time to freshen up the ad through more testing and newer material.

Once you begin to get traffic and sales, you can measure your performance. There are two main things we want to accomplish with A/B Testing. One is to measure the clickthrough rate of each ad. The other is to measure the conversion ratio. The conversion ratio is the number of sales divided by the number of people who clicked on your ad times 100, or Cr = S/Ct*100. This will give a percentage.

Conversion ratios vary widely from business to business, so it is impossible to say here just what you can expect. You might be able to get a range by asking at some of the business forums. Give it a try.

This sounds like a lot of work (it is) and a lot of expense. But remember, as your ads get more successful, your income should more than pay for continued testing. It is a cost of doing business. Without it, you will only be relying on blind luck to succeed. It might happen for someone, but that someone won't be you. Use good business practices and stay with it. Success will come.

-0-

Will Moore is the owner and author of a number of websites, including the highly successful Web Stats Gold. www.webstatsgold.com provides high power e-commerce analysis for small and medium sized e-commerce web sites. Contact Will at will@webstatsgold.com.

Copyright © August 2004 by WG Moore. All rights reserved

You may reprint this article so long as it is not modified in any way and includes both the author bio and copyright notices, including the link back to Web Stats Gold.