Multi Variate Testing – How to Use Multi Variate Testing to Boost Website Conversions

Transcribed from Video:

Charlie Hutton: Good morning, everybody. It’s Charlie here from Hutchinson Web Design and today, we are going to take a look at Multi Variate Testing and how it can be used to help your business to make sure that it’s website is performing at its absolute maximum when it comes to converting your visitors to leads, customers or general inquiries.

Now I know we recently looked at split testing. Now the difference with split testing is that you can only really one solid bit of information versus another bit of information. Multi Variate Testing really is the crème de la crème when it comes to testing your website for performance and it means you can test and measure hundreds and hundreds of different options all at the same time without having to really think about it. You can let it run over X number of days, come back and the whole system will tell you which is the best combination so that you are sure you’re getting the best conversions or the best return on your investment.

So let’s dive in and let’s have a quick look at how Multi Variate Testing works in the field. We’ve got a sample page set up here with an opt-in form on. This is probably a squeeze page. Now the goal of this page is to get people to click this “send” button here. So if we were doing a multivariate test on this webpage, the goal of the test would be to see which or what we need to do to make sure what combination of options we need to have to get the most conversions or visitors to click through to this “send” button.

So if we were doing this sort of thing, there’s probably four different blocks here that we would test or that we could change. The first one here, we might want to change the title of the webpage. We might want to change or test this image here. We might also want to test this phrase on the signup form at the top of the form to entice people and we might also want to test the actual wording used in the button. So really, if we were looking at testing this page, we would be testing four blocks. We would have block one, block two, block three and block four.

Multi variate testing

Now with multivariate rate testing, essentially for each one of these blocks, what we would do is we would assign a different option to each one. So for example, for the title, we could have three different titles that we might want to try. They might be using different words or different key phrases or different sales phrases that we might want to test out to see which one is going to perform best. Our image for example, we might have three or four different images that we want to test.

In this case, the first one might be with clouds. It might have another image which has got a sun on it or another image which is showing rain; but basically, by testing the image out, we can see which image or which sorts of image have the most impact with the visitors to our website and which image the visitors relate to most of all when it comes to clicking this signup button here.

The same goes again for headline on our form. We could have four different headlines here. It could be something like “Please Enter Your Details,” “Enter Your Details Now,” or “Start Now and Save.” So we could have three or four different headline options that we want to test to see which one is performing and the same might go again for our “send” button here.

We might have had an argument in the office of whether we should have a red or a green button or maybe whether the button should say “send” or “submit” or “order”. I mean there could be loads of different options here but basically we want to see which ones or which combination is going to get the most clickthroughs and the most conversions to our website. Obviously the more conversions we have, the more leads that we can have in our office system that could be followed up by sales staff or via email. More leads generally leads to more revenue if they’re done properly. So they’re really the reason behind this Multi Variate Testing is to make sure that our websites are optimized 100 percent so we can generate more revenue for our business.

So with these blocks in place on this website, what the Multi Variate Testing does is we would say right, OK, we’ve set the goal. The goal is to get people to click this “send” button. We would then let’s run this test for a thousand visitors by using a bit of clever scripting and code. What we can do is each time a visitor comes to our website, the software will automatically pull a different option  through for each one of these blocks.

So visitor one to the website will see a completely different page to what visitor two might see. So visitor one for example might get tied to one here. They might have an image of the sun here. They might have headline two here and they might have a green button with a “submit” where visitor two, they might have title three. They might have the image of a night sky. They might have a headline that says “Fill out details below”. They might have a red button and it might say “send”.

Now basically, as I said, every time a different visitor visits the webpage, they will see a different combination of these options. Now the Multi Variate Testing, what that does is once these sales and visitors has been reached, it will tell us or it will pull through and tell you which is the best combination of these different options to use to get the maximum conversion rate on people clicking the “send” button. So each time that it tries a different combination of these options, it will record whether or not someone clicked through to the “send”, whether it went through to the goal or not. Each time it goes through to the goal, it will record that combination and give it sort of like a plus one.

So once we know the conversion rate of this page, we can do some really, really simple math and some figures. Work out exactly what our cost per lead is and then see how that affects everything in the whole grand scale of things.

Let’s say before, you were roughly converting your website at 20 percent so of every hundred visitors that went to your website, roughly 20 of them will become your leads. Now this was before any Multi Variate Testing was done. So let’s say that you’re sending traffic to your website using Google AdWords and each click costs you one pound. So if we sent 100 visitors, our cost for those 100 visitors would be 100 pounds.

Working on the 20 percent here with 100 visitors, roughly 20 of those would become leads. Let’s say in this example we know that on average, every lead that we get through via our website, we’ll spend around five pounds. So from those 20 leads, we’ve each roughly spent five pounds and that gives us a revenue generated from our pay per click campaign of 100 pounds. We then got to minus the hundred pounds that we spent to send traffic to our website and you can see we’re only converting at 20 percent in this example. We’re just breaking even on those figures.

So, as I said, when sitting down and working out the numbers and what you do and what you’re spending on your AdWords or your pay per click or your SEO campaigns to get traffic to your website. You need to be working on this conversion rate number to make sure that you’re getting the best value for your money or the best return in your investment.

Let’s say that using our Multi Variate Testing, we increase that 20 percent conversion to a 40 percent conversion rate. Now, for the same 100 pounds that we spent on pay per click, rather than 20 leads, that’s actually double now and we get 40 leads. Again, we still know that our average spend hasn’t changed per lead because we know that they’re going to spend five pounds or on average of these 40, that we spend five pounds.

So that gives us a turnover for a 40 percent conversion rate of 200 pounds. We then minus our cost of the campaign which is 100 pounds and what we have there is we have then made 100 pounds on top of what we spent originally. So basically, once we know this figure here or once we know what our webpage converts are – and for all intensive purposes, it could be a lot higher than 40 percent. If we know what our webpage converts are, we know what roughly the average spend per lead is that comes through from our website. We can send traffic to that website all day long.

If we’re converting at 40 percent and we know it costs us 100 pounds to get 40 leads in; and we know from those 40 leads we’re going to generate 200 pounds, for every hundred pound that I spend, I know I am going to make 100 pounds in profit. It’s an absolute no-brainer. I can sit there and I would sit there and I would do sit there and spend 100 pounds on pay per click advertising every single day because I know for every 100 pounds that I would be spending, once I know this 40 percent figure, I know this five-pound figure, I know exactly what I’m going to be making at the end of the day. Well, in this scenario I would be able to spend that 100 pounds quick enough because I knew that for every 100 pounds that I would be spending, there would be an extra 100 pounds in my back pocket or in my business’s back pocket.

Multi Variate Testing Summary

So as you can see, Multi Variate Testing. Once you know your figures, obviously these numbers will be scaled dependent on the size of your business and what your venture is. I imagine for a lot of you, this five-pound figure is minimal but it just makes the numbers easy for what we’re going through here. You can really see by doing the Multi Variate Testing and doing the optimization on your website, you can make a huge difference from actually making a loss on your traffic-driving campaigns or breaking even to actually getting in profit. As I said, once you know these figures, it really is a license to print money.

So if you’re not doing it already, start doing some Multi Variate Testing or speak to us about Multi Variate Testing and we can make sure that your landing pages, your squeeze pages and your sales pages are maximized for conversions so you guys know those numbers and can start making some money on the traffic that you’re driving to your website.

So thanks very much for listening. It’s Charlie from Hutchinson Web Design and we look forward to seeing you soon.


Multi Variate Testing


Leave a Reply

Your email address will not be published.