Newb Guide Part 7: Split Testing

This is part 7 of Nickycakes’ Affiliate Marketing Tutorial.  Please read the previous parts if this confuses you.

Once you have some successful, or close to successful ad campaigns running, you will likely want to start refining those campaigns to maximize your profits.  Since, as mentioned earlier in this guide, you know jack shit about marketing, your only real option for doing this is by testing.  Split testing is simply the process of making two variations of part of your campaign, be it your ad, landing page, affiliate offer, whatever, and splitting your traffic between the two to find which works better.

Split Testing Ads

Split testing ads is easy with most of the popular PPC platforms such as Adwords, MSN Adcenter, Yahoo, etc.  The basic idea is, have 2 separate sets of ad copy running at all times, and periodically get rid of the one that does worse, and make a new one.  Some people (like nickycakes) do this as often as once a day, and some much less frequently, or not at all.

When starting out, it may be good to make several different ad copies, maybe like 5 or so, but make sure they are all very different.  After running the ad for a while, you can pick out the winner, and start split testing that ad by making very small variations.  Seriously, like only change ONE thing about the ad.  There are so many things to choose from, sentence structure, caps, no caps, display url, etc.  Just change one thing each time you’re split testing ads.

How do you know when the number of impressions the ads have gotten is enough to be statistically significant so you can make a proper decision?  Easy, just use this split testing calculator (thanks dr ngo): here.

Split Testing Landing Pages

This is a little more involved than ad copy split testing, but it’s the same idea.  Split traffic between two different landing pages to see which one converts better.  Your goals will probably be something like reducing bounce rate, increasing clickthrough rate to the offer, and increasing overall conversion rate on the offer.  To do this you will likely need a decent analytics package.  Here are some you may want to check out:

ClickTale ( – This one is a little pricy, but if you have the cash, it’s apparently well worth it.  It tracks literally everything you could want to track on a page.  Where people are clicking, how long they’re taking to enter information in a form field, whatever.

-CrazyEgg ( – Much less expensive, but doesn’t quite track as much.  You probably want to look into this one if you don’t feel like shelling out a lot of money but want some solid analytics.

-ClickHeat - A free heatmap package.  Not too comprehensive.

-Google Analytics – Decent features and free, but not realtime.  If you’re cool with only being able to see your stats once a day, and not having that many useful stats to look at, its better than nothing.

-Google Website Optimizer – Ok, Nickycakes has never seen this one in action, but apparently it’s pretty good for doing split/multivariate testing on landing pages.  If anyone has some experience to share, please leave it in the comments.

Split Testing Offers

This one is extremely important.  There are so many similar offers out there that it’s impossible to tell which one is going to convert best with your landing page.  IGNORE the EPC data that the networks give you for the offers.  They are often meaningless bullshit, and will have no bearing on how well it will convert for you.  They could have had some idiot direct linking ppv traffic to the best offer on the network driving the EPC down to like 1/10th of what it will perform for you.  So, as usual, test for yourself.  Just like with ad copy split testing, assuming there are several choices for similar products, split test 2 of them, wait till you have enough data to determine which is the better offer, drop the underperformer, and split test the winner with a new offer.  Pretty simple.  If you have enough traffic you could, of course, just split it between all of them at once and find out the winner in a day.

Split Testing Call of Duty 4 Guns

So, which SMG is better? The P90 or the MP5?  Try 5 rounds with the P90 and then 5 rounds with the MP5 and see which one gets you more kills and headshots.  Obviously the mp5 is going to win hands down unless you’re spray-and-pray newbslice, but again, it’s good to test for yourself.

Next part will be about taking over the world.

Keep it real.

Peanut Gallery

Add a new comment