Learn the Basics of Testing Ad Copy: A PPC Primer

By ,

9 SHARES

One of the best features of pay-per-click advertising is the ability to test. There are numerous tools and options in AdWords, Yahoo! Search Marketing and adCenter that allow PPC advertisers to test nearly every piece of their account. But, admittedly, this barrage of bells and whistles could make the average-Joe-PPC-beginner go crazy – or worse yet, abandon PPC altogether! I would recommend to any advertiser interested in testing their PPC variables to begin with ad texts. As such, today’s PPC primer is on learning the basics of testing ad copy.

Why Should You Test Ad Copy?

Many of you reading this post will already know the answer to this question, but it never hurts to have a reminder. Testing ad copy will allow you to incrementally experiment with variations in your messaging to determine which ad texts have the highest click-through-rate (CTR) and/or conversion rate. This process is often referred to as A/B Testing or “split testing” and is extremely important to the overall health and growth of a PPC account.

How to Get the Most Accurate Results

Ensuring that your PPC account is operating with the correct settings is very important to obtaining accurate results from your testing. What this means in terms of testing ad copy is that you must have your ad delivery options set to rotate evenly (all ads in a test will accrue the same amount of impressions). In Google AdWords, this is quite simple. Enter one of your campaigns and click on “Edit Campaign Settings.” From here, scroll down to the “Advanced Settings” section and click the radio button next to “Rotate.”

adwordsrotate.JPG

This setting will now rotate your ads evenly over time. The alternative is to let Google determine the best ads in your ad groups by CTR alone – which takes all control of an accurate test away from you!

Yahoo! has a similar feature with their “Optimize Ads” option. To get accurate testing results: enter one of your campaigns, select all of your ad groups and click the “Optimize Ad Display” button. This will bring up a dialogue box with the options of saying yes or no to optimization. Just say no.

yahoooptimize.JPG

This is the simplest way to adjust the Yahoo!’s ad delivery options. However, you can also access this feature through each ad group’s settings.

MSN’s adCenter is a step behind when it comes to testing ad copy. At this time, adCenter automatically optimizes your ads based on CTR. There is not an option to turn this off.

Over time, adCenter will automatically display the better performing ads more frequently, which can help boost the performance of your ad group.

How to Get Started

Let’s assume that you only have a single ad in your ad groups – this will become your control ad with which to test against. Use the control ad as your starting point when writing new ads. If I can make one point clear it would be this: test one change in ad copy at a time. If you want to update the headline, change only the headline. If you want to update a call-to-action in line 2 of the description, change only the call-to-action, etc. Making incremental changes will allow you to quickly make the right decisions based on the outcome of your split testing. In this example, I am testing the effectiveness of my headline:

adexample.JPG

What Should You Test in Your Ad Copy?

The list of ad text variables is quite extensive and can vary dramatically depending on the product, service or niche an ad is written for. To get you thinking about what kinds of things to test, here is a great list from the AdWords blog:

  • Different emphasis: product description, call-to-action, or promotional offer
  • Including the brand name versus simply describing the offering
  • Including the price in ad text versus including a discount or other special offer to differentiate your business
  • Including an audience-specific message such as ‘Perfect for Couples’
  • Placement of certain messages in your ad text: headline, line 2, or line 3

When to Determine a Winning Ad

As a rule of thumb, I typically wait until both ads (or however many ads are being tested) have all generated at least 30 clicks. This is to ensure that each ad has had equal opportunity to be displayed and will therefore provide accurate data. For those who are more statistically inclined, there are other options. For one, you can use the Teasley Statistical Calculator which will give a much more accurate depiction of your split testing endeavors.

When you’ve determined a winner in the testing of your ad copy, it isn’t time to quit and move on. No sir – it’s time to write a new ad and test another variable in your ad text! Make testing ad copy a priority on your PPC management task list. Keep up-to-date on your tests and consistently try to out-do yourself in terms of creativity and trying to think like your customers. Good luck, and happy testing!

Get more weekly links with our Fast Five newsletter! Five Fast Links in Your Email Every Friday.

Also send me a daily RSS digest

Social Advertising Toolkit

Twitter Facebook LinkedIn Google+ Email Print More
  • http://www.portentinteractive.com michaelportent

    “MSN’s adCenter is a step behind when it comes to testing ad copy.”

    adCenter is a step behind on everything. But that’s another article, right? =P

    Another important thing when testing ad copy (I think) is to keep the total number of variations in each ad group between 3-4 unless you’re getting thousands upon thousands of impressions/clicks per day. For lower-traffic campaigns, if you test 7-8 ad variations, none of them will get enough looks in a week to make conclusive judgments about which one(s) were most effective.

  • Pingback: (EMP) E-Marketing Performance » : » Team Reading List 1.30.08

  • http://www.ppchero.com John

    Michael,

    An adCenter rant article probably isn’t too far down the road…

    You’re absolutely correct about the number of ads to test at one time. Our team recently inherited an account from another firm and in practically every ad group there were 6 or 7 ads that had no clear purpose or direction for testing. It was a blatant reminder of the age old adage: KISS. Keep it Simple Stupid! Testing ad copy is important, but not at the sake of your sanity. : )

  • Pingback: The Ultimate List of PPC Ad Testing Resources | The ClickEquations Blog

  • Pingback: Team Reading List 1.30.08 | Pole Position Marketing

  • http://twitter.com/j_camarena Julio Camarena

    If you want to test out your ads, to turn on the rotation is a must, google indeed optimize for a better ad, bust most of the time they don’t give a fair chance to “prove” which ad gets better results .. a human can do that far better.

  • Akeila

    How long should you test ad copy to determine which ad performs the best?

    • http://www.hanapinmarketing.com PPC Hero

      You should run your test until you have significantly significant data. Make sure that you have enough clicks/impressions to know which ad is winning. We use the Teasley Calculator around here (http://www.teasley.net/calculator.html).

  • Jerome hambrick

    Great article John…as a new copywriter…this is valuable information.

  • Rich

    Hello,

    When split testing, how do you account for positioning? Naturally the lower you are position wise CTR tends to follow the downward trend. If two ads have different overall ad positions after the test is complete where one has a better position than the other is that still considered a viable test?

    Thanks!
    Rich

    • http://www.hanapinmarketing.com PPC Hero

      Hi Rich,

      Based on the way the auction is run your ads are all going to have the same bids. It’ll be on either the keyword or ad group level, but never on the ad level. Because of that setup your ads should theoretically have the same average position.

      However, as you’ve probably noticed, that’s not always true. An ad that does better in terms of average position is one that Google’s systems seem to prefer (since it’s more competitive in the auction). For our split tests we assume that average position is a wash based on the way the bids work, but if you see big differences in average position you will be able to tell which ad Google likes more, which can then help you decide how to conclude your test. Hope that makes sense!

      • Rich

        Thank you so much!!!