AB Testing

A Beginner’s Guide to A/B Testing PPC Ads: Part 1

Pay per click ads are a worthwhile investment for many online marketing campaigns, but they can also end up being a waste of your advertising budget if they aren’t used effectively. If people are clicking on your ad but immediately bouncing away from the page it takes them to, you’re losing out on both money and conversions.

You can’t just create a PPC ad and then sit back, assuming that it’s going to be effective. The good news, though, is that because pay per clicks are directly measurable, you can continually analyze your data using A/B testing and make adjustments in order to get the best possible ROI.

What Is A/B Testing?

You may already be familiar with A/B testing if you’ve had previous online marketing or web design experience, but if it’s all new to you, here’s a brief overview.

A/B testing is an either-or test where you compare two web pages or ads that are identical except for one variable. This is done by randomly assigning traffic to each ad and seeing which one results in more click-throughs, time spent on the landing page, and conversions. For example, if you’re testing a PPC ad, you might look at how tweaking the headline slightly affects your click-through rate.

The most important thing to remember with A/B testing is that you can only test one variable at a time (as you would do in any good empirical experiment). If you get carried away and change multiple aspects of your ad, you’ll be unable to tell which factor is responsible for your results.

So how do you decide what to test first? Read on to learn how to systematically test your PPC ads.

Define Your Goal

The very first thing you need to do is determine what’s more important to you: click-throughs or conversions.

If the main goal of your advertising campaign is to gain greater exposure and develop your brand name, then click-throughs will probably be the most important thing to you. If you’re selling a product that consumers typically spend more time researching, click-through rates may also be valuable to you, because they show that potential customers are gathering information on your site that will hopefully inform their buying decision.

If the main goal of your advertising campaign is to inspire some kind of direct action, such as getting site visitors to sign up for your newsletter or purchase one of your products, then you’ll be more interested in looking at actual conversion rates. This is what most businesses are going to be interested in; after all, if you’re spending money on your ads, you want to make sure they’re giving you a good ROI.

Choose Your Variable

There are typically 4 main variables that you’ll want to test:

  • Headline: The short, to-the-point line that viewers will usually click to visit the link
  • Links: The landing page or product page that a viewer is taken to when they click the headline
  • Body text: The brief description of what the viewer will get if they click through
  • Keywords: The words that Internet users search for that will lead them to your ad

You’ll want to test all 4 of these variables, but choose just one to start with. Let’s say that you choose the headline. Take your ad and come up with at least 3 different versions of the headline that you want to try.

Wait and Watch

Test at least 3 ads at the same time for the same variable. Depending on the amount of traffic you get, you’ll want to run the ads anywhere from a few days to a month to get the most accurate read. Usually 2 to 3 weeks is a good time frame, but if your results aren’t clear at this point or if they are conflicting (maybe one version has a high click-through rate but the other has a much higher conversion rate), run the test for longer. Sometimes results can be complicated, so you may want to consider using an online A/B testing tool like Optimizely or Google Analytics to get a better sense of what’s going on.

Once you do get clear results, cycle out the lowest performer and run the test again using a different variable.

What to Keep in Mind

Make sure you’re running ad variations at the same time. If you run one version of your ad, then wait and run another, time is another variable you have to account for. Let’s say you release one ad closer to the holidays: that one might have a higher click-through rate because more people are looking for gifts on your site, but the original ad may have actually been more effective. By running ads simultaneously, you don’t need to worry about time as a variable.

Results may contradict your instincts. As William Faulkner said, “In writing, you must kill all your darlings.” The principle holds true with advertising as well. There may be one particular ad that you’re really attached to, or a headline that you think is going to instantly convert thousands of web users, but if it’s not performing well in A/B testing, you should phase it out. Remember, A/B testing is letting you look at real results from real consumers, so pay attention to the data.

Pay attention to where your ads land. Where your ad takes web users is just as, if not more, important than the content of the ad itself. You may have crafted a killer ad that makes tons of people want to learn more about your brand, but if they click through to a web page that’s confusing or convoluted or looks like it hasn’t been updated since the ‘90s, they’re likely to click away pretty quickly.

Set up an optimized landing page. Rather than just having each ad link to your home page, set up an optimized landing page. Keep in mind that you can also A/B test variables on this page to find out which version is getting the most conversions.

Now that you know some more about how A/B testing works, you’ll want to start thinking about how to actually craft the best possible ad for your money. I’ll be covering all that in the coming weeks, and next week I’ll look specifically at how to write the most effective headline.

1 thought on “A Beginner’s Guide to A/B Testing PPC Ads: Part 1”

Leave a Comment

Your email address will not be published. Required fields are marked *

Click to access the login or register cheese