The quality of your ads is decisive when determining the quality factor and, thus, the competitive capacity of your keywords in the context of an AdWords auction. Split tests provide one measure to continuously improve your ad quality. But which ones are the ads that perform the best? The majority of PPC managers consider the average click-through rate – adjusted for the ad’s position – to be the most important criterion.
In certain cases, however, it has proved useful to go into further detail. What are the search terms behind these average values? Is ad A really getting placed for the same search terms as ad B, or are we comparing apples and oranges here?
The AdWords client center allows you to analyse your ads’ performance on search term level in order to find answers to the questions above. You don’t know how to go about this? I’ll show you…
Step 1: Create and download your AdWords search term report
Most of you probably already know the AdWords search term report to be a useful tool for identifying new or negative keywords. Only few people know: you can extend the report to include information about your ads.
- Open the search term report (e.g. tab Dimensions > view: Search Terms).
- Set filters and columns as required.
- Click on “Download report” (see graph below).
- Select Segment > Add segment > Ad.
- You may find it useful to select “Network (with search partners)” and “Device” as further segments.
- Download the report.
Basically, that’s all there is to it. Unfortunately, Google doesn’t let us off the hook this easily. At this stage, all you can see in your report are the columns of ad status, ad type, Display URL, and Destination URL.
Step 2: Indicate headline and description lines in your search term report
In order to have your report display also other ad attributes, such as, for example, headline or description lines, we need to take a detour via the ad ID.
As outlined above, we again download a search term report. However, this time, we select “Ad ID” for our segment. By clicking on the tab “Ads”, we can now download a report that also includes information about the Ad ID (adjust columns, download report). Now, we need to merge the two reports using an appropriate tool. Note that the Ad ID is only unambiguous in combination with the Ad Group ID. Since the latter is not included in the report, you need to work with the name of the ad group instead. If you are confident to clear all the CSV-import hurdles and if you are processing only few data, you may choose to work with Excel (see VLOOKUP).
Step 3: The elegant way: AdWords Script
If you need this report frequently, we recommend that you refer to the reporting feature of AdWords Script. It allows you to download the search term performance report, to link it via Creative ID and Ad Group ID to the desired ad data, and to enter the results, for example, into a Google spreadsheet. Once you have created the script, next time, this will automatically work at the click of a button.
Further possible applications: Mobile Preferred Ads, keyword cannibalisation
The report as outlined above may also prove useful in areas other than your typical ad A/B test. For example, it enables you to analyse for which search terms Google places Mobile Preferred Ads and for which search terms these really perform the best.
Furthermore, the search term report illuminates the topic of keyword cannibalisation from a whole new angle: For which search terms and how frequently does cannibalisation, in fact, lead to the user being shown a semantically less specific ad?
The catch: Only search terms with clicks
There is one major shortcoming involved in using the search term report. Google only releases information on ads that generate at least one impression, but not necessarily clicks. Accordingly, the comparison of ads on search term level is missing valuable information.
Hence, the search term report will never replace conventional ad tests. Nevertheless, it’s a useful addition to have, especially in the area of high traffic.
Now that the steps I have outlined above have laid the foundations for an evaluation, I will provide a real-life analysis of the combination “Ad performance via search query” in my next blog entry. Stay tuned!