Simple A/B Testing for Pay Per Click Ad Text Elements
Published March 31, 2008 by Garry Przyklenk PPC Basics
In my sole post last week, I mentioned how you might get into searchers heads by finding and understanding personas for your ad text, landing pages, and conversion stream. This week, I’m going to focus on how search marketers can setup matrices to test and fine tune specific aspects of their ad text to maximize clicks to landing pages. Fine tuning click through rate is probably the easiest thing you can do to help conversion, here’s why…
Testing and fine tuning your ad text maximizes your click through rates for high-quality clicks, meaning visitors are more likely to convert through a purchase, registration, or something else you want them to do.
Setting up a simple matrix to test the effectiveness of new ad text is simple. In fact, a matrix can be as small as only two variates (versions) of an ad. The only real-estate you have control over in your ad is:
- Headline text
- Description text (usually two lines)
- Display URL
Altering any of the text in any of these spaces will result in changes in click through rate (CTR). The most substantial changes usually occur when modifying your headline text, as this is the text most search users notice first and foremost. Headline text can make or break your ad in terms of CTR, or it can pre-qualify visitors that aren’t serious in purchasing just yet.
Description fields are usually much longer than headlines and can make your job much harder. You do have a lot of extra space, but what will make your ad get noticed, and what will make your ad convert? Testing is the only way to find out.
Display URLs are a little more complicated, because Google Adwords has new rules about what can be used as a display URL depending on your target URL. However, it is a modifiable field and can be tweaked for conversion.
How should you test?
Having a bachelor of science isn’t required, thankfully, but since I have one I can give you a few suggestions. Whe testing ads:
- Keep a control ad, one ad that is kept unchanged for each and every test.
- Make sure you choose “rotate ads” so that each ad gets equal billing.
- After writing ads, make sure all of them are actually running (you might get disapproved ads after submission to your search engine)
- Before declaring a “winner” run a statistical significance test to determine error.
What should you test?
Different search marketers will test different things they deem appropriate. I don’t think the search marketing game is that different from other marketing efforts, advertisers do bring experience, creativity, and opinion to the table. Your best course of action is to test outcomes that will prove or reject variations based on past experience, minimizing options as you go.
Here are some elements in your text ads to play with:
- “(BRAND) Official Site” versus Keyword-only headlines
- Price listed in the ad versus prices on the landing page only
- Free shipping or not (depending on your product)
- Superlatives like “best” or “cheapest” are golden, if you can get exceptions for them (proof you are the best)
- “www.” versus not having “www.” before your domain name
- Directories after your display URL versus not having directories
- Using “Free” anywhere in your ad versus no mention
- Dynamic keyword insertion
Keep in mind that with pay per click, you do not necessarily have to wait very long to find a winner. As long as you ensure your results are statistically significantly, and once you test one element of an ad versus a control ad, you can start building a matrix for your results. You may find that ads mentioning your brand, price, and using “www.” give you the best click through rates for quality traffic that converts best.
You can’t claim to know what works for sure until you test. And don’t become complacent over time because factors outside of our control can change the outcome of ads and conversion.
