My First Foray Into Split Testing
Over the last few years I’ve been studying internet marketing, but only recently have I started to look seriously at it. I’m in the process of setting up a couple of different subscription based membership websites, and the first steps in building these websites is to develop a sales funnel to start collecting leads. For those of you who are totally green, a lead is basically someone who you have determined to be a potential customer. How do you do this? You find people who are willing to provide you with their contact details (phone number, email, address etc.) in return for something of interest to them. You could be providing them with some information that they want, a training video of some kind, or in an offline example it could be a free product sample.
So the next question is, why is it useful to know the contact details of someone that wanted a free sample of acai berries, or a free training video that shows them how to improve their golf swing? Well because it is much more likely that this person is will to buy more acai berries, or to buy an extended golf training DVD than someone who didn’t even want to enter their contact information. However, just because someone wants a free sample of what you’re offering doesn’t mean that they are going to buy from you, in fact it’s quite likely that only a small percentage of the leads you collect will ever end up buying from you. Anyway, that’s enough about leads to give you an idea of what I’m talking about, I’ll go into more depth on this whole process another time.
The first part of my sales funnel was to build what is often referred to as a landing page, or email squeeze page where I try to entice people to provide me with their email address in return for more information on the product that I am developing. Now here’s the thing, I haven’t completed any free offering yet. I don’t have a free sample, a related product I can give away, nothing. All I have is the promise that I will give them more information on my website once it is complete. This is not very enticing, which means that my opt in rate (percentage of people who come to the page who actually provide their details) won’t be as high as it could be, actually much lower. Nonetheless I can begin testing to see what I can do to improve my opt in rate.
How do I test? Testing is basically a statistical process where you drive a certain amount of traffic to your landing page (in my case via pay per click advertising), and track how many people opt in. Then you make changes to your landing page, and once again you track how many people opt in. If you send 1000 people to each version of your landing page, and version A has an opt in rate of 3% while version B has an opt in rate of 4.5%, then you can say that version B is a better page, you have proof of it. It is best to test with as much volume as possible, to insure the most statistical accuracy, anything below 1000 people and you might have a difficult time determining if it’s a true trend, or just an anomaly.
I’m using Google Adwords to drive traffic to my landing page, and while I haven’t seen anywhere near 1000 visitors yet I have tweaked a couple things based on the recommendation of some people in forums. Now that I have implemented some fundamental lessons that I learned, I plan to allow a decent amount of traffic to go through to the site in order to see how well it performs. After 1000 or so visitors I will make some changes and see what difference it makes.
When I talk about the changes I will make to the site I am referring to the sales copy. I’m sure a lot of you have read some cheesy looking sales pages on the internet and thought to yourself that nobody would ever buy from this page, but you’d be wrong. These marketers are generally testing their sales copy in the same way that I described above, and they can show you statistical proof that these pages DO WORK, not to mention the money they are making.
By determining what your opt in rate is you are finding a piece of the puzzle that will help you to determine how much you will have to spend on customer acquisition. To give you a little more information, by determining for example that 4.5% of all visitors to your landing page provide your with their contact details, and 10% of those people end up buying your product (based on your follow up marketing efforts) which has a margin of $50.00 per unit, you now know that if you sent 10,000 people to your landing page that 45 of them would end up buying from you, providing you with a margin of $2250. That means that you need to spend less than $2250 in order to drive 10,000 visitors to your landing page, or to go even further than that each visitor to your website is worth no more than $0.225 ($2250/10,000=$0.225 but we’ll round down to $0.22), and each lead is worth $5.00 ($50*10% purchase rate=$5.00 per lead)
I apologize if I haven’t done the greatest job of illustrating all of this to you, I have a long ways to go before I master all of these concepts. I understand them in theory, but I have yet to fully execute a lot of them at this point in time. I’ll chronicle my progress as best I can without revealing too many specifics about my individual projects.