In PPC (and most other forms of marketing), the name of the game is "testing." If you don’t test different variables within your campaign, you will never be sure that the results you are getting are the absolute best. But what types of testing should you be doing? And what things should you be testing? Below are some of the basic testing strategies used in PPC that advertisers use to refine their campaigns and improve performance.
There are two main types of tests that PPC advertisers will run in their campaigns – A/B testing and Multivariate testing:
- A/B Testing (aka A/B Split Testing) – These types of tests are normally the starting point for advertisers looking to refine their campaigns. Basically, an A/B test involves the implementation of a single change to whatever element you are looking to test, and comparing that against the original version of that element. An example here would be changing an image on a landing page, or switching the headline of an ad copy variation. A/B tests are easy to create, and are ideal for smaller campaigns. The major downside to A/B testing, however, is that by only testing one variable at a time, finding a winning version can take a long time.
- Multivariate Testing -With multivariate testing, you are testing multiple variables within a single element at the same time. For example, if you were testing two separate landing pages, one using headline A and image A, and the other using headline B and image B, you are carrying out a multivariate test. These types of tests allow you to try out many different variables in a short amount of time. The more complex the multivariate test, however, the more difficult it can become to determine a clear winning combination.
Once you’ve determined which type of test you will be conducting, you will need to choose some elements to test. Below are a few common test areas for PPC advertisers:
- Landing Page – Your landing pages tend to be the most tested components of your PPC campaign, and for good reason – this is the element that will ultimately drive (or turn away) a user toward a conversion. Practically everything on your landing pages can be tested, and depending on whether you are carrying out an A/B or Mulitvariate test, you might be testing several elements on your landing pages at once. Some of the more common landing page elements that are tested include title tags, H1 tags, calls to action, button/form placement, images, colors, text length and format, and navigation structure.
- Ad Copy – Within your PPC account itself, ad copy is likely to be something that is tested on a regular basis. (If it isn’t, it should be!) Everything from headlines to body copy to display and destination URLs should be tested. The only time you should ever stop testing ad copy is when you’ve reached a conversion rate of 100% – and since that is not likely to happen, ad copy testing should be a regular part of your on-going PPC campaign maintenance.
- Keywords – From bids to match types, your PPC keywords offer up ample testing opportunities. For brand new campaigns and ad groups, one strategy you can implement is including multiple match types of the same keyword. Once you are able to gather some data and see which match type performs best for that keyword, you can pause the other match types. Along the same lines of this strategy, you can also adjust your keyword bids per match type to see how it affects performance. Generally, I will set the exact match version as the highest bid, the phrase match about 15-20% less than the exact match bid, and the broad match about 35-40% less than the exact match bid.
Experiment with different testing types and testing elements. At the end of the day, as long as you are continuing to test things in your campaigns, your PPC performance will improve.