Better Conversion Tip #9 – Test
Split testing, also known as A/B testing, is the process of having two identical pages (except for one change) rotate randomly to see which one generates more sales (or opt ins, or whatever the goal of the landing page).
A little piece of code called a “cookie” is stored on their computer, so if they return to your page again, they will always see the same page where they originally landed. This is important because you don’t want them coming to your landing page and seeing one deal, then returning later and seeing another.
Whenever you contemplate making a change, you should always split test the change against the control (the control page is the one that is generating the most sales or opt ins or whatever). The object of split testing is to always try to beat your control. Then the one that beat it becomes the new control, and the process starts again.
If you take away anything from this tip about split testing, it should be this: Split testing is an ongoing process. It never ends. You should always try to beat your control. Sometimes one small change will make a big difference. This is especially true when split testing headlines, because more than 5 times as many people read your headline than they do your copy. So headlines are a good start to get the biggest bang for your efforts.
In fact…
…although the tips and tactics in the previous articles in this series are based on real-world results for most markets, they should still be tested. Not all markets react the same way to all things.
Remember that people vote with their wallets. It doesn’t matter what they say or do. The end result is how well your landing page is performing. Period.
Now how do we know which page is the “winner?” Well, you might expect that’s whichever one produces the highest conversion rate.
In most cases, yes, assuming all things are equal. For example, if I’m testing the headline, and one headline produces a better conversion than the other, I can usually assume the headline with the higher conversion will become my new control to beat.
But if I am testing the price or offer, things can be a little different.
I could, in fact, get a better conversion rate with landing page “B” over landing page “A.” But if this is a test of the price, then B could produce a better conversion rate, but I’ll actually make less money, because of the cheaper price in B. And because in many cases you’ll sell more items when you lower the price, then what really is the right price in which to sell the item?
Well, obviously you need to make a profit, so at a minimum your price will have to be high enough to cover your expenses and then some. But how much of a profit should I set for each item? Answer….whichever will maximise revenue and that will make you the most money.
You might be thinking that it should be the highest that the market will bear. And my answer above does take that into consideration. But it’s really just a little more complex than that.
You see, certainly it’s possible to sell more items at one price and make less money than fewer items sold at a higher price. And don’t forget you have to factor in return rates and support costs as well. It may very well be more profitable for you to sell a handful at a much higher price than to sell many for cheaper.
The best way to figure out what you can sell your product for and make the most in profits is to split test your price and offer until you find that “sweet spot” (and it will be sweet when the money starts flowing in!).
That way you can do a complete end-to-end analysis of your costs, profits, and support overhead for each price point until you find the magic number.
So that brings me back to my original question. What is the best way to determine which landing page “wins” the split tests. The answer is your return on investment (ROI).
Conversion rate by itself becomes a useless figure when you are comparing apples to oranges, like in a price or offer split test. But ROI? It’s a true measure of a page’s effectiveness.
Because then you know for every dollar you spend, you’re going to get X in return from page A and Y in return from page B. And if Y is greater than X, well then page A is your new control, the winner.
Now, here are some of the elements you should test on all of your landing pages:
- Your headline (most likely to produce the biggest difference)
- Your lead (the copy right after the headline)
- Your offer (what they are getting as a whole for the price)
- Your price
- Your P.S.’s (one of the most read parts of your landing page after the headline)
- Whether to include audio or video
- Whether the audio should start automatically or only when the visitor initiates it
- Different payment plans
- Whether to include a photo of you or not
- Individual captions under the photos
- Your layout
- Your fonts and colors
- Whether or not to include a header graphic
- Different types of header graphics
- Whether to use a signature or a type of signature (here’s a great place to get computer generated signatures: http://www.vletter.com)
- Whether to add a toll free or non-toll free phone number
- And much more (basically any change you are considering that could have an impact on the goal of your landing page)
The headline and the offer are the items that will probably make the most difference so start your testing with these.
As a general rule, once a landing page you are testing gets at least 30 desired actions (sales, opt ins, etc.), the winner is the one with the highest ROI during the test. Then you move onto the next test and start over again with the winner as your new control to beat.
This one process alone will likely cause tremendous improvements for all your landing pages.