Thread: Are Your Maximizing Your PPC Split Testing Strategies?

Split testing, also known as A/B testing, is commonplace for most pay-per-click (PPC) advertisers. The idea of testing two different ad copies or landing pages is nothing new, but I bet there are many advertisers that could dive a little deeper when it comes to split testing and discover a new level of clarity for their PPC campaigns.

First, here is a quick definition of split testing as regards PPC. With split testing, two or more items are tested side by side to determine which performs better at a specific metric. The importance of testing is critical to optimization. If you don’t test, you cannot improve your campaign. It is as simple as that. Most advertisers will find that the first couple of tests will have the most dramatic results, but it is important to continue testing various elements of your campaign at all times.

So what can you test when it comes to PPC? I think the most prominent — and therefore important — thing to test is ad copy. Testing multiple ads for every keyword you are bidding on is essential to PPC success. Testing new headlines, description lines, display URLs, and destination URLs are all important areas to test when it comes to ad copy split testing. The idea of testing multiple ad copy translates very well to the content network also. Rich media campaigns are probably on the forefront of split testing and, if they are a component of your PPC campaign, you need to make sure to test new creative all the time.

On ad copy level, split testing should run for a minimum of one week, but running for one month is ideal. Every day has different characteristics, and each week has variances (think about pay weeks). Ads should attain a minimum of 300 clicks, but 1,000 is ideal. You need to allow for enough data to accumulate for each ad before making any decisions regarding how effective it is. You also want to let every ad gather at least 15 conversions. Each advertiser will have a different set of results. This is simply a rough guideline.

Landing page testing is another very important aspect to test when it comes to any online marketing campaign, PPC included. Testing new layouts, copy, fonts, colors, forms, and shopping carts are just a few things to consider when split testing your landing pages. Knowing how much weight your landing page has in regards to quality score and conversion rates, you should spend amply time testing out countless versions throughout your campaign.

A few other things to make sure to split test are match type, ad position, max CPC, time of day, day of week, and month of year. Look at it this way, anything you have an option for should be tested. This also applies to testing elements. Clickthrough rate, conversion rate, total conversions, cost per conversion, page views per click, cost per click, and total cost are just a few outcomes that most PPC campaigns test.

Make sure you take that deep dive when it comes to split testing within your PPC campaign. It is a large ocean and even though surfing the waves may be appealing, you don’t want to miss out on what lies beneath the surface.

Hi,

A smarter approach recommended by Stompernet is to create 4 identical copies of your control ad, and one test ad. This ensures that your “safe” ad will show 80% of the time, and your test ad only 20%. Plus, the control ad’s solid click through history will cause it to rank at higher positions than brand new ads, so creating new copies reduces this bias.