Posted by James Lee on 1st May 2012

How to get the most from Ad Copy Testing

If you want to develop pay per click ad copy that delivers results, a sound testing strategy that compares the performance of multiple ad copy versions is essential. However, when setting out to evaluate or test different ad copy versions, it is also important to understand ad serving methodology as this can dramatically affect your results.

Recent improvements by Google have resulted in three options for ad serving:

  1. Show ads with high click through rates more frequently
  2. Show ads with the highest conversion rates more frequently
  3. Show ad versions equally regardless of “early” conversion or click through rates

If you observe the timing when Google starts serving more ads to one ad copy version over others, you will notice that it often occurs before sufficient data has been generated.  If ad serving is skewed in favor of one ad copy version, then the ad copy evaluation may also be skewed.

To illustrate how the basis for ad copy serving can impact results, I have presented a real case study. In this case study, 2 ad copy versions were tested in 2 different time periods producing very different results.

Case 1.

Ad copy                                % Served             CTR                        Conv                      Cost/Conv

Copy A                                 84.2%                    2.64%                    11                           $30

Copy B                                  15.8%                    2.51%                    6                              $10.9

Case 2.

Copy A                                59%                        2.16%                    12                           $16.4

Copy B                                40%                        2.62%                    4                              $39


Based on this data, which ad copy do you believe has the best conversion performance?

Chances are you selected Copy B from Case 1 and Copy A from Case 2 as the winner of ad copy testing in terms of conversion performance.

If you observe the data closely, you can see Case 2 presents convincing data that copy A is superior to copy B.  Why is this not true for Case 1?  The reason is that Copy A’s served percentage in Case 1 is 5 times greater than Copy B’s.  As you can see, the ad served percentage had a significant impact on results.

From this data, we can conclude the following:

  1. Ad copy performance may not be scalable.   The inconsistent performance of Copy A (Case 1 vs. Case 2) suggests that ad copy performance is not scalable as Copy A is served 5 times more than Copy B.
  2. Google’s click based ad copy serving method may not the best option when profitability is the objective. The latest trend in Google’s ad serving logic would be to display ad copy B more often based on its higher click through rate even though ad copy A shows superior conversion performance and overall profit generation.
  3. The best approach for evaluating ad copy is to distribute ads evenly until there is sufficient performance data to make an informed decision. As illustrated in the scenario above, ad copy A’s performance was superior to copy B in terms of maximum profit generation.  However, when ad distribution is uneven, performance data may be inconclusive or skewed, potentially leading to the wrong decision on ad copy performance.


In conclusion, to select the best ad copy and optimize ad performance, it is not only important to test copy versions but also to ensure that each version is served evenly for a sufficient period of time in order to generate true profit driven performance data.  Once a sufficient amount of performance data is generated, it is then necessary to establish the performance difference that will distinguish winning ad copy from copy that will perform poorly.  Don’t fall into the trap of continuing to run multiple ad copy versions and poor performing ads long after you have the necessary data to tell you which ads and ad copy will produce the best results.

To assist advertisers with the process of testing and determining the highest performing ad copy, ClickSweeper provides a profit driven approach to ad copy evaluation. ClickSweeper’s profit driven approach determines ad quality based on click through rate and cost per conversion. Once the data is in and poor performing ad copy has been identified, the poor performing ads can automatically be suspended in each Adgroup so that you can stop wasting money and diluting your efforts by continuing to run underperforming ads.

For More Information

No comments yet!

Post your comments