The AdProof Blog

12 Feb2013

A/B Testing Google AdWords


Many online marketing campaigns regard pay-per-click (PPC) advertising as a vital component in its strategy. Apart from being amongst the most significant factors, it can also be the most expensive, which is why it’s vital to keep track of all the investments.

The first step towards leading an effective PPC campaign is to decide what to write in an advertisement. In order for the advertisement to work in our interest, there are four key factors we need to concentrate on. These consist of:

The Headline

The headline is the first line of the advertisement, and it must be strong enough to grasp the readers’ attention quickly. In this case, shorter is better. The headline is there to shine light upon what is being offered. An ideal length would be 4-5 words, but the aim here is to keep your message as short and to-the-point as possible. The headline will be in a different color than the rest of the text, usually blue. The headline may also consist of keywords that will show up in bold in a Google search.

The Body Text

The body text will consist of 1-2 lines, and should be similar to a page’s meta-tag description. Searchers will be looking for something that will give them a little more idea of what you have to offer, and your body text must be strong enough to convey the message.


Take some time to think of what terms a consumer might search on to find your product or company. At this point, testing on what keywords work best and coming up with keyword-advertisement combinations will help you succeed.

The Landing Page

The landing page also pays an important part in PPC advertisements. The URL, in order to be effective, should take the consumer to the company’s product page, unless the company’s homepage acts second to a product page.

Tip: Consider camel casing each word and using your target keyword in the display URL to increase click through rate. Example:

In this example, we’ll use Google AdWords to demonstrate proper PPC A/B testing or split testing. In order for the campaign to work to its maximum potential, the user must show variations by split-testing advertisements. Most users will not have access to complex tools, but low-tech variations can be run from something as simple as an Excel spreadsheet.

Tip: At the initial stage, dynamic insertion should be ignored and ad optimization should be turned off, since we want all advertisements to be shown with equal probability.

Assuming that only the advertisement and not the landing page is being tested; you can come up with a table that looks something like this:

Ad Title Description 1 Description 2
1 1 1 1
2 1 1 2
3 1 2 1
4 1 2 2
5 2 1 1
6 2 1 2
7 2 2 1
8 2 2 2

Following exponential numbers, we just created 8 advertisements with 2 variations on each line. As the number of variations increase, the numbers go up, and this is where the testing comes in. As far as the testing goes, there are some important factors we need to keep in mind. These consist of:

  • All variations must be tested simultaneously to reduce time-based factors that may result in biased results.
  • Test one thing at a time, so you can tell what’s affecting the success of your advertisements.
  • Let the numbers decide what’s right, not your instincts.
  • Test your advertisements long enough. For example, depending upon the type of product the company has to offer, the advertisement may get less than 100 impressions a day, or 10,000 impressions a week. In any case, you must have a few thousand impressions to reach an unbiased conclusion. Also, any special events like holidays coming up should be taken into account, since these usually result in a different type of traffic.

Keep in mind that advertisements and landing pages will be directly related to each other, and different advertisement and landing page combinations can lead to different conversion rates.

Results can also be customized to show you what you’re looking for. Custom reports can help you organize your reports by CTR, number and rate of transactions, and total profits. Reports like transaction rates also help you keep track of how many consumers are returning customers.

A/B testing may seem like a lot of work, but it will prove beneficial in the long run, since it helps you get the most out of your advertising budget. Having access to advanced toolsets will definitely help, but low-scale split-testing may prove just as good.

Have you signed up for AdProof, yet? AdProof lets you A/B test your ads in minutes. AdProof makes A/B testing PPC ads a walk in the park and costs less than a cup of coffee. Sign up and get your first ad test for free.