Independent Activity 2: SplitTester: An Example

Paying for ads on Google, Yahoo or MSN (called Pay Per Click or PPC) can cost a lot of money, but if done well can generate a lot of sales. Typically, people using PPC run 2 or more ads and split test them to see which one is the better performer. Once they find the best they try another ad and test again until they have a very good performing ad. The Click Through Rate (CTR) measures performance. The more people click on your ad, the better its performance. However, this entire testing can eats up a lot of advertising dollars. SplitTester is a free tool, which helps choose the better performing ad based on only a small number of ads

All About SplitTester
SplitTester is a clever little tool that gives you the real lowdown on a mathematical principle called "Statistical Significance." Here's what that means:Let's say you flip a coin five times, and voila! it comes up heads each time. Does that mean it will always happen that way? Of course not. You just got lucky. If you do the 5-flip test 32 times, you'd expect naturally to get all heads at least once. That's just the laws of statistics and probability.

Now here's a helpful little rule of thumb for you: If you're running two ads and trying to compare their performance, by the time you have 30 or more clicks or actions you've pretty much established how those ads will perform against each other, long-term. But waiting for multiple ads to reach 30 clicks or actions can sometimes take a long time. Sometimes you need to know now how two ads will perform against each other without having to wait until they've had a large number of clicks.

That's where SplitTester comes in. SplitTester can take smaller numbers of clicks and do the statistical calculations for you, telling you how confident you can or can't be that your two ads will have different long-term response rates. It will tell you whether you're on the right track or not with your ad campaigns.Just enter the number of clicks and the click-through rates for each of your ads, and SplitTester will do the rest.

Go to SplitTester and test the following ads in the two different case studies to see if one ad is more likely to be a better performer in each scenario

Case study 1 –
Ad A gets 20 clicks with a CTR of 2%
Ad B gets 20 clicks with a CTR of 2.5%

Can you be confidant that Ad B is the better performer (higher CTR)?

Result: You are not very confident that the ads will have any different long-term response rates.

Cast study 2 –
Ad A gets 20 clicks with a CTR of 2%

Ad B gets 25 clicks with a CTR of 3%

Can you be confidant that Ad B is the better performer (more clicks, even higher CTR)?

Result: You are approximately 90% confident that ad B will have better long-term response rates.

StatCounter | Other Tools | Split Testing | SplitTester: An Example