If you’re not A/B testing ad copy, you’re probably missing out on revenue. But with so many different aspects of a simple search ad to play with, how should we know where to focus?
This post explores several areas of ad copy A/B testing that sometimes fly under the radar in favor of more commonly tested ad components. Make no mistake, there is opportunity to improve ROI , especially for the meticulous search engine marketer.
Your job as ad copy tester is to map out all the possible benefits and technical tweaks you can enrich your ads with, and test all of them in a highly methodical manner. Let’s walk through some of the elements that have been shown to make an impact.
What Should I Be Testing?
I’ll go into detail on a few of these, but first I want to give you the lay of the land.
- Dynamic keyword insertion
- Competitive landscape
- Qualifiers, e.g. price in ad copy
- Punctuation
- Speaking to your audience
When testing, the key is to stay focused and process-oriented. Here’s an effective process to follow:
- Start with big ad concepts
- Work your way down from concepts to the details
- Give each test enough time to demonstrate statistical significance
- Don’t test too many things simultaneously in one ad
- Measure everything you possibly can
- Push offline conversion data back into AdWords and optimize ads based on it
Now, let’s dive into some testable components of AdWords ad copy and look at examples of demonstrated success… and failure.
Testing Dynamic Keyword Insertion
This is a feature that’s easy to understand but tricky to master. Here’s an explanation of how dynamic keyword insertion (DKI) works.
Novice marketers sometimes jump to using DKI because it seems like an effortless way to make your ad appear to offer exactly what your visitor is searching for. But it takes more than that to show serious results with DKI.
In the early days of PPC, heavyweight search players like Amazon and eBay were criticized for over using DKI. You’d search for nouns like “honesty” or “integrity” and right away, eBay’s ad would pop up encouraging you to “Find Integrity at eBay.” They’ve since toned down and sharpened their game, but you still find examples of how DKI can go wrong:
Screenshot taken Jan. 21, 2014.
What is a coffee shade, exactly? Is it available at your local grocery store?
Mmmm. Coffee.
Image from en.wikipedia.org, labeled for reuse.
These ads can still be effective, but don’t let your brand get caught with its proverbial “dynamic keyword pants” down. It provides a poor experience for searchers.
On the other end of the spectrum, PPC Associates’ Todd Mintz talks about an interesting DKI success story in a WordStream survey post about A/B testing surprises. In one campaign targeting hundreds of brands in a merchant’s vertical, he tested a static ad headline against a dynamic one:
50% Off Widgets
50% Off {Keyword: Brand}
Obviously, the ad headline has been changed to protect the client, but you get the idea. Using specific brand names dynamically in the ad headline took this campaign from being an unnoticed, nearly zero conversion campaign to a triumphant perch as the top-performing campaign in the entire account. As Mintz puts it, “I thought something broke in AdWords.”
The lesson is, if you’re careful about how you use DKI and continue to test new ideas, it can quickly become a winning strategy — but it’s not an “easy answer” to writing strong ad copy.
Some keys to good DKI usage include:
- Test DKI against similar ads without DKI (don’t assume it’ll perform better)
- Conduct copious test searches to ensure nothing looks weird
- Use very small, tightly themed adgroups, i.e. one or two super-relevant keywords
- Use a highly relevant default keyword (which appears if the search phrase is too long for your headline, and this happens often)
Does Punctuation And Other Minutia Matter?
The short answer: Yes. The long answer: It depends. If you’re just gearing up to start a major ad testing program, leave punctuation and small details of your ad copy until later. You first need to nail down which sweeping ad concepts work better to attract the visitors you desire.
When you’re at this level of optimization, though, it’s important to remember that with only 100 or so characters to differentiate you from the next guy, every little morsel can be important.
Consider this case study from a test run by the team at Wordstream, where four different instances of ad copy happened to use exclamation points. The caveat here is that this wasn’t, strictly speaking, a test of exclamation point against the lack of one (with the exact same copy), but it does indicate a trend. It suggests that punctuation matters.
It can even make a difference where you decide to place a comma. Perry Marshall discusses this A/B test, where the ad variant on top showed an 8 percent CTR improvement by inserting a single comma into the headline.
Screenshot taken Jan. 21, 2014.
Also consider how your ad will appear when it lands in the first few positions in the highlighted box, where instead of two lines of description, you have just one. Will the punctuation you’re using stand up to this format?
This LegalMatch ad, for instance, is clearly written for the sidebar format. Notice how the word “Free” looks redundant?
Screenshot taken Jan. 21, 2014.
If it were in the sidebar format, it would look much more palatable:
Find a Lawyer – Free
Free, Confidential Lawyer Locator
Save Time – Describe Your Case Now!
The trick is to write ads that work in both formats.
An online coffee bean retailer writes an ad that looks like this in the Google sidebar:
Screenshot taken Jan. 21, 2014.
When this ad scores a top spot, the way it’s written stands up to the format:
Screenshot taken Jan. 21, 2014.
Speaking to Your Audience
Another critical piece of those 175 allotted AdWords characters is who you’re speaking to. Before writing your ad, double-check that you understand who your audience is — and once you’ve written something, check again for whether your copy speaks to that audience.
Take this example, where Wordstream’s aim was to sell signage to businesses who need to remain compliant with the Americans with Disabilities Act (ADA).
Ad #1 below focuses on the boldness and clarity of the signs themselves — which are great benefits — but this ad loses because the buyers of the signage are not the readers of the signs to whom clarity might be appealing. The buyers are businesses, and they care about the benefits in Ad #2: Avoiding fines and staying compliant with employment law.
Screenshot taken Jan. 21, 2014.
The winning ad goes farther and actually calls out business owners in the copy, too. The increase in CTR was 164 percent. Similarly, when Amplify Interactive shifted ad focus from marketing managers to webmasters and web developers in AdWords campaigns promoting a SaaS marketing tool, the result was a 152 percent lift in sign ups and a 58 percent cost reduction. The lesson here is in building personas and focusing on the real target audience.
Ask a nearby co-worker to imagine they are the audience you’re targeting — name the persona specifically, i.e. imagine you’re a barista, or a circus performer — and challenge them to tell you whether the ad you just wrote speaks to them.
So far, we’ve touched on several key areas of SEM ad copy testing — but many others lurk. As your A/B tests become more advanced, consider testing these elements, too:
- Value Proposition/Benefits Testing
- Headlines
- Description Copy
- Display URL
- Sitelinks
- Capitalization
It’s a never-ending story, really. This is an exercise for the detail-oriented. In some of these case studies, though, you can see how a tiny change like the placement of a comma can yield dramatic improvement in conversion rate. Best practices are good, but testing within your real-life AdWords account with ads and keywords specific to your business provides valuable insight that generalized advice simply cannot.
You should feel good as long as you have a number of tests running at any given time — if that’s happening, rest assured, you’re improving your campaigns!
Do you have an interesting AdWords A/B test that supports or contradicts something in this post? I’d love to hear your thoughts in the comments.
