If you’re not split testing almost everything you do, you are either making no progress, or leaving way too much money on the table in your online business. All great marketers and many great businesses do this, and it can literally take you from terrible results to great results faster than you would imagine.
Split testing can be done with ads, sales letters, product names, pricing, bonuses, communication methods, and almost anything that can be tracked for some sort of effectiveness. I’m going to talk about split testing with Google Adwords, wgere it’s easy to implement, and the results can be dramatic. Then, I’ll touch briefly on split testing other parts of your business.
For every AdGroup you are running in Google AdWords, you should always be running 2 different ads. Some people recommend 3, but that actually confuses things, mathematically speaking. Just be sure to set your campaign setting so rotate ads evenly (NOT optimized for the better ad), and create a second ad for each AdGroup, in which you change ONE THING about the ad – the headline, for example, or adding the word ‘Guaranteed’.
Do NOT change too many things at once, or you won’t be able to tell which change made the difference, which is the whole point of split testing. So, if you want to test headlines, change ONLY the headline. Got it? Good.
Normally, you’re going to need about 30 or more clicks on each ad before you can say which ad is really performing better. However, there’s a great online tool that will help you make this decision with fewer clicks, compliments of Perry Marshall and Brian Teasley:
Just enter the number of clicks and click-through rate for each of the competing ads, and this tool will tell you the statistical probability that one ad will perform better over the long run. You might also end up with a tie, in which case you should just try another ad.
If you go through this process on a regular basis, and make enough changes to your ads, while tracking everything along the way, you will eventually end up with the BEST POSSIBLE ADS for your campaign, because they just keep getting better and better! Wouldn’t you want to increase the click-through rates and sales from your ads by a factor of 10%, or 50%, or 100%, or 300%?!
The answer is YES!!!
The same holds true for your sales letters. Great copywriters are always shooting to write the next “control.” That’s industry jargon for the current best sales piece out there. All revisions to mailers, brochures, etc. are tested against the control on a small scale first. If they beat the control in testing, they’re rolled out to the market at large, and if they still outperform the control, they become the new control.
You can do the same. You can buy software, or even use Google AdWords tools to setup sales pages or landing pages on your site that alternate from one to the other, while you keep track of the sales or opt-ins from each one. Once you have that data, keep the good one and write another variation, trying to beat the control.
Believe it or not, product prices control sales results, and not always in a way you would think would be obvious. Sometimes, raising the price of a product will make you more money.
That’s why you should TEST EVERYTHING.
If someone tries to tell you their OPINION about how you should market something, first ask them for data to back up their opinion. If they have some, great. If they don’t, consider testing their ideas, but realize there is no truth without testing the results for yourself.
In fact, even if they do have data to support their claim, they probably have a different list, different product, and other differences that should probably make you just test everything anyway. Ideally, you want all parts of your sales process optimized for your product, your business, your communication style, etc.
So you thought once you got out of school, the testing days were over? Sorry, if you want to be successful, your testing days have only just begun!
If you have any testing stories to share, or if you have any comments on this post, please post them in the comments below. If you like this post, please ReTweet it through your Twitter account, using the ReTweet button at the top of the post.
Hey you're kinda new here aren't you... You should subscribe to our RSS feed, so you never miss a post. It's also my mission to get to know all my new readers. Follow me on Twitter so we can chat. Thanks for visiting!
Lee from handmade...
{ 2 comments… read them below or add one }
Split testing is a good idea for your add campaigns, but I have a question for. Do you think it’s a good idea to do split testing when it comes to your actual website as well?
Yes, but don’t look to close at my own websites!
It’s a good idea to split-test virtually everything you do in your business, IMHO. In fact, the correct answer to “Is this a ‘good’ headline/webpage/sales letter/whatever?” is “I have no idea – you’ll just have to test it.”
Even marketing experts don’t know the answer. They have some pretty good ideas about what will *probably* work better than most things, but the only true measure is a real-life test.