Most PPC and affiliate guides recommend split testing. I agree split testing is one of the main keys to success. But the only thing they usually mention is split testing of ads. The truth is almost every aspect of a campaign can and should be split tested to optimize efficiency and profits. Testing is not just something you do at the start of a campaign, but a never ending process.
Here are some things I like to split test.
- Domains – I have talked before about the importance of testing multiple domains in your campaigns. The display URL is a critical part of your ad, and you never know what will convert better until you split test different domains.
- Offers/Networks- The same offers are usually found on several different affiliate networks. By split testing to the same offer on different networks you might find a higher conversion rate on a particular network.
- Account structure - Try taking the same set of keywords on 2 different accounts with a different adgroup structure to see which performs better.
- Keyword match type – Try testing all available match types to see what converts better.
- Landing Pages – Landing pages are critical to split test. You can almost always find a better converting landing page. You can split test any page change down to something as simple as button colors.
- Ads - Of course ads are the easiest thing to split test, and still one of the most effective ways t0 optimize your campaigns.
These are just a few, I’m sure there are a ton of other things to split test. What creative ways do you split test?
We can’t stress enough to our clients how important testing is. You may be loosing out a lot if you’re not testing your ads.
But only split test one thing at a time or you wont be able to tell what is causing the changes
Not completely true Neil, sure if you’re manually (ick boring) testing by hand you’ll need to test everything individually.
However split testing software has really come of age and some (like the Evolutionary Splittester) can test every different component (like headline, subheading, price etc) of a landing/sales page and tell you which component works the best to create the perfect sales page.
Best of all, it does it all automatically, no more boring change headline, wait a month, change again, wait a month etc (man I wasted so many hours of my life doing that…)
Great post Chad!
I’ve always wondered why an offer might perform better on one network than another (assuming the LPs are identical). Never could figure that out, but I’ve seen it many times.
I split test offers as well from different networks. Same offer on one network will perform better than the other. I could never understand that. Also some advertisers are notorious when it comes to shaving leads. So I run the same offer from Network A for certain time of the day and the switch it to the same offer from Network B thereafter.
I wouldn’t want to rely on third party software to split test my stuff. Sure I’ll get all these metrics and stats but doing with hand one at a time like Neil suggested and going slow is more comfortable for me. Of course when I have 50+ campaign going already I may need to look at automation tools like that.
niche-affiliate-marketing.com
Chad – Assuming there is no lead shaving going on, and assuming you are sending traffic to the same landing page (your own – i.e., not direct linking), can you explain why the same offer may convert better on one network over another?
[...] was necessary to touch on each one, however, after reading a post by Chad from CDFnetworks entitled Split Test Everything, I decided to go ahead and continue doing articles on split [...]
What’s your reasoning for not using an automated tool Gee? Just trying to understand why you’d want to waste time doing something that can be automated?
Cheers, Tim
Just curious – are you using Adwords A/B Tester or your own solution to split test?
Nothing wrong with using automated tool. In fact my job is to create automate tools to make lives easier at my job. I’m a bit slow goer. If something is working for me I tend to stick to for longer even if when it takes me longer. I don’t want to disturb the rythem. Like I said, when I want to launch 10+ campaign at one time, I will definitely make use of automation tools.
No I don’t A/B tester. I create at least two ads for each adgroup with at least two landing pages (different in design and content) for testing. Let it run for few days before have some data to see which combination ads, landing pages are working. I keep tweaking them (both ads and landing pages) until I’m happy with the return. I do that this one campaign at a time.
Hope that was helpful. you can read more on my blog.
niche-affiliate-marketing.com
[...] was necessary to touch on each one, however, after reading a post by Chad from CDFnetworks entitled Split Test Everything, I decided to go ahead and continue doing articles on split [...]
[...] fact so have a ton of other bloggers like Chad from CDF Networks in which he wrote a post entitled Split Test Everything! Jonathan Volk wrote a post on Increasing Landing Page Conversions where he talks about split [...]
[...] fact so have a ton of other bloggers like Chad from CDF Networks in which he wrote a post entitled Split Test Everything! Jonathan Volk wrote a post on Increasing Landing Page Conversions where he talks about split [...]