I was glued to the computer monitor watching the real-time results come in. No, this wasn’t an election, this was business and the stakes were high. I’d convinced my biggest client at that time to make a big change and I was eager to see the results. I was absolutely convinced that the results would be good, but until the results were in, the question would remain: Was this the right thing to do or was this just a waste of time and money? What happened next remains vividly imprinted on my mind more than a decade later.
It all started with a question as we discussed the sales letter that he himself had written. “Is it possible that a sales letter with a different approach would bring in a much greater response?” “How do you know that if you use this sales letter, you’re not leaving a lot of money on the table?” With these questions I had his attention. The only way to know for sure was to test it. After some discussion he reluctantly agreed, and with that, preparation got underway for split testing a sales letter to a mailing list almost half a million strong.
As I refreshed my computer screen and studied the real-time results that were already starting to accumulate from the sales letter, my eyes almost popped out of my head – the sales letter that I had written outperformed the other by a margin of more than 4 to 1. That morning my client became a firm believer in the power of split testing as a powerful tool to increase the response rate of advertising.
Introduction to split testing
Split testing is a method of direct response advertising in which two versions of an advertisement are run instead of just one. The goal is to determine which version of the advertisement produces a greater response rate. In order for the response to be tracked accurately, the response must be able to be tied back to the original ad. For direct mail, this is done with a unique department code that is assigned to each ad. For local businesses, this is done with a coupon or even different bonus incentive offers for each ad version. For Internet pay-per-click (PPC) advertising, the “click rate” and “conversion rate” are compared between ad versions.
The winning ad is the version that achieves the highest response rate in a test. Sometimes the results will be dramatically different and at other times the two ads will achieve very similar response rates.
Online then Offline Advertising
One of my favorite offline advertising strategies is to begin the ad campaign online and then move it offline – AFTER we’ve run through the first several iterations of split testing. Online advertising campaigns can be deployed, tweaked, and re-deployed so much faster than offline advertising. For this reason, it is much cheaper and faster to do enough split test advertising online that you are comfortable with the response rates you’re getting – before moving to offline advertising. I love testing several advertising headlines online with search or social media advertising and then using the strongest ones in print.
This combination approach provides you with the best of both advertising worlds. With your online advertising you’re able to ensure the faster evolution of your advertising campaign, and with your offline advertising you’re able to reach your target audience in a way that continues to be effective, even in a digital age.
Rinse and repeat
The result of split testing an advertisement is that you learn what achieves the best response rate with your audience at that time. The real objective here is to constantly improve your response rates so that your ads are consistently achieving the highest possible responses. This happens by discarding the losing ad, keeping the winning ad, and creating a new variation of the winning ad for your next test. Your next ad campaign will contain the winner of your last ad split test and a new ad. Think of this as advertising evolution. Instead of survival of the fittest, you have survival of the winningest.
How can you make your advertising make more money? Test every ad. Only then will you be sure that your advertising is achieving the highest response rates.
How quickly can you put split testing to work for you and your business?