Wondering where to begin with ad copy testing? Columnist Jason Puckett shares some ideas to get you started.
One of the most important attributes of a successful PPC manager is the ability to incorporate new technologies in existing management strategies. With the official sunset of AdWords standard text ads and the introduction of AdWords IF Functions, we must develop ways to fuse the new with the old.
At my company, we see how today’s top advertisers are using these changing technologies to evolve their creative testing strategies. This article is designed to keep you up to date and fuel your creative testing fire with 10 testing ideas that are sure to drive huge performance changes.
Best practices for ad copy testing
Before we get started with testing ideas, let’s touch on good fundamentals.
Aggregate data across ad groups and campaigns (where it makes sense). The more volume, the better. Aggregating data across similar ad groups or campaigns can help you speed up your tests and uncover more reliable findings. There are technologies out there that will help you improve the specificity of your ad creative at the ad group level, while still testing at scale. Ad Customizers and IF Functions are two great resources.
Even if a test doesn’t yield a winner, still consider it a win. Sometimes testers get discouraged if they don’t encounter a big win in their first test iterations. Remember, proving something is a non-winner is just as important. These insights will help you prevent making costly mistakes in the future.
Multivariate test to find the truest results. If you don’t test all possible permutations within an experiment, your results may be invalid. Remember, some creative may only work well when paired with other creative elements or landing pages — or some creative may only work well within certain campaigns. These differences make multivariate testing imperative to any good testing strategy.
While many articles have been written about effective ad copy, the truth is that winning ad creative will vary depending on a variety of factors. Thus, it’s critical to test for yourself to determine what works best for your business. Following are some ideas on which elements to test.
1. Numerical abbreviations vs. full numerical values
Using different manipulations of numerical values is a method you can use to grab attention and differentiate your ads. Flaunting your inventory, listing numbers, product usage or customer base is a great way to cut through the noise and instantly establish credibility.
2. Promotional quantification method
Knowing how your customers respond to your promotional efforts is absolutely necessary to an optimized advertising campaign. We have seen clients drive drastic changes in click-through rates and sales figures by adjusting how they approach their promotional language.
Wording and promotional types will vary by industry, but here is an example of three ways to advertise during the same promotion. Please note the Headline 1:
3. The first word (verb) in your call to action
No matter which industry you are in, a portion of users will view ads as threatening, some will perceive ads as helpful, and others will believe they aren’t relevant to their decision-making process. A big difference-maker in the interpretation of your digital presence is how you’re asking the user to engage with your brand.
The first word in your call to action (CTA) can make or break someone’s encounter with your brand. Optimize that first impression by testing the first word in your CTA.
4. The placement of your call to action
Depending on the category of your business and the stage of the buying process a searcher is in, you may need to vary the placement of your call to action. This is a perfect example of why multivariate testing is so important.
A multivariate test allows you to uncover the optimal placement of your CTA for each audience individually (done with the proper segmentation). Try using your CTA in Headline 2 and at the end of your Description.
5. Use your brand name in Headline 2 for your non-brand keywords/campaigns
Even if your brand isn’t in the Fortune 500, including brand name in the Headline 2 of your non-brand campaigns can be quite helpful. Branding your ads helps to develop a persona, so the user knows they aren’t clicking on a random link but rather a company they can trust. Warm, fuzzy brand sentiment is always a good thing; if you’re successful, maybe you’ll earn repeat business from this customer.
6. Multiple descriptions that use AdWords IF functions
Evolving your A/B testing methods to incorporate new technologies is vastly important. Think about writing description variations that contain both a mobile and desktop version. You won’t be able to test which device message is being shown/clicked (AdWords reporting constraint), but you can create several descriptions, each of which is optimized for audience or device, and measure which variation works best as a whole.
7. Headlines that use ad customizers
I love this as a test, and as a best practice. If a portion of your search Headlines doesn’t contain a piece of dynamically inserted content, you are probably leaving money on the table for two reasons:
- Your Headlines probably aren’t specific enough to your ad groups, or
- They aren’t flexible enough to implement ad copy tests across campaigns/ad groups.
In the image below, each product SKU in the inventory would be inserted into Headline 1; surrounding ad copy can be changed and tested. Try testing variations that do and do not implement this practice.
8. Dynamic keyword insertion (DKI) vs. no DKI
Our studies have shown that DKI does not improve click-through rate as a whole when compared to well-written ads (as shown in a similar study here). If you’re still using DKI and are fearful of implementing other strategies that may cause a slight decrease in performance, testing is a good way to offset this risk.
Set your ad rotation to “auto-optimize for clicks.” This will allow your existing ads to gain the majority of the initial impression share. But as the new variations begin to outperform your baseline, the sample will become more evenly distributed. Before you comment, “An even sample is the only way to test,” it isn’t. If you need to offset some risk, avoiding “rotate indefinitely” is a great way to do that.
9. Landing page testing
If your landing pages are set at the ad level, and you have created two separate landing pages that you’d like to test, you can use the same ad with both destination URLs.
If you’re feeling ambitious, try using multiple variations of ad creative — for example, a multivariate test with two landing pages and two descriptions (four total variations) as shown below:
10. Benefit testing in your Headline 2
As referenced in a previous article, Headline 2 is a great way to grab attention. One way of catching a searcher’s eye is using benefit-based language. Rather than talking about features, prices or promos, advertisers should use ad copy that is going to evoke emotion from the reader. Explaining exactly how your product or service is going to make readers’ lives easier is a great way to evoke that emotion and the click you desire.
Test multiple variations of benefit-based language. Think about your value propositions, and turn them into ad copy. Your test should be set up in a way that allows you to extract which benefit users respond to most positively.
Stay on top of your testing
Testing is a great way to improve account performance, but it should be conducted with caution. After you uncover a winning ad variation, be sure to implement it as soon as possible in order to “cash in” on that improvement. Also, keep an eye on under-performing “non-winners.” As soon as you can designate a variation as a non-winner, eliminate it from the account. If you let your under-performing ads run for too long, overall performance can be pulled down.
Please let us know if you have any questions. We’d like to hear from you.
Some opinions expressed in this article may be those of a guest author and not necessarily Marketing Land. Staff authors are listed here.