The ABCs of A/B testing.
An old saying in advertising by John Wanamaker is that half the money spent on advertising is wasted; we just don’t know which half. Ummm, no thanks. We don’t have 50% of our advertising to waste in entrepreneurship. I love paid advertising because it can show you measurable results, clear ROI attribution, and, most importantly, the ability to test.
Whenever you create an ad strategy, there are elements based on industry expertise, market research, past performance and other factors. These inputs are used to develop a working hypothesis on ad strategy, but it needs to be proven with any theory. Factors to consider include how knowledgeable the market is about your product or your brand? What is your price strategy? What is the length of your buying cycle? For all of the planning, when launching a paid ads plan, we get to put our hypothesis on the audience, targeting, messaging, imagery and all the elements that make up an ad into the market to see real results. This process always gets my blood pumping, and then the real fun begins…we get to learn!
The length of that learning curve depends on many of the factors earlier listed, and the best way to refine learning is through testing. This process compares different variations of your ad composition or strategy to see what works best and what can be improved. The need to test is reasonably understood, but there are certain best practices for testing to help you gather more precise and more consistent results. Here they are:
TEST ONE THING AT A TIME.
When conducting any type of test, you want to isolate a variable to know what is impacting your results. Variables could include, spend, time of day your ads are shown, image, copy, or any other element that makes up your ad. If you start changing many factors, it is difficult to understand which variable is responsible for positive or negative impacts on your results. Change one thing at a time, and you will quickly refine your performance and have more conclusive results.
CREATE A MEASURABLE ITERATION TO YOUR HYPOTHESIS.
When testing your ads, you want to create a measurable iteration to your hypothesis that you can tangibly interpret. An example of this could be “Do I get better results if I show my ads seven days a week or only on weekdays?” or ”Will my cost per lead change if I add more budget to this campaign?” You can then start to compare results and interpret any changes with a clear understanding of what caused any changes.
USE ONE IDEAL AUDIENCE FOR THE TEST.
Sometimes people want to test on a sample audience size before expanding to a larger audience. It is important that an audience is large enough to support the test so that the results aren’t irrelevant once expanded. It would help if you also didn’t use this audience for other campaigns at the same time; overlapping audiences can cause other delivery problems for your campaigns.
SET THE RIGHT BUDGET.
Similarly to the audience, it is important to test with a budget that will produce enough results to determine a winning strategy. This one can be challenging as every penny counts in an advertising campaign and we always want those dollars to contribute to ROI. To that end, it is always important to set a budget you are comfortable with, but testing must be done so that the results support positive ROI for the program in the future.
USE AN IDEAL TIMEFRAME.
For me, this is the big one! When you adjust a variable in your ad strategy, you need to give it time to gather your results. Depending on your customer’s buying cycle, a minimum of 7 days is the recommended time frame for a test. If your buying journey is longer than that, you should give enough time for your specific conversion activity to occur.
The beginning weeks of a paid ads program are exciting but also nerve-wracking. For any business every dollar is important, but when you are a small business or entrepreneur, the stakes feel even higher with when and how the budget is spent. So, if sales don’t come pouring in the first few days of starting your ads program, take a breath and go to the data. Do you have an intelligible working hypothesis? Do you have the right audience size and budget? If so, let your plan run for at least that first week before making changes. Then, change one thing at a time to test and refine your working hypothesis into proven results.
One of the beautiful things about paid ads is that the numbers can tell you a story and guide you to adjust to achieve your goals. If something isn’t working, absolutely make changes to it, but do it in a thoughtful and precise way so that you can learn, grow, and achieve results. As we like to say, stick to the plan, keep your arms and legs inside the ride at all times, and most importantly, have fun!
