The 5 principles of testing

Our entire philosophy is based on the principle of testing EVERYTHING you do — and tracking the results.

It’s the only way to keep track of what you’re doing right with your online business, and what you need to improve.

For example, let’s say you want to make some changes to the layout of your web site. Obviously, you don’t want to make changes just for the heck of it. Any changes you make should be for one main purpose: to do a better job of converting visitors into customers.

But if you don’t test the new version vs. the old version, how will you know if the changes you make actually DO improve the performance of your website? And if you don’t track the results of your testing, how will you be able to know exactly which changes had the greatest effect?

Simply put, if you don’t test and track every single change you make to your site… every promotion you run… and every traffic generation strategy you try… you’re going to be shooting in the dark every time. And there’s absolutely no way of knowing what you’re going to hit.

So here’s my list of the top 5 things you absolutely MUST know about testing to keep your business on target.

1. Only test ONE thing at a time

The most important thing to remember is this: KEEP IT SIMPLE.

Test one thing at a time. It’s the only way you’ll ever know exactly which change is causing which effect.

Let’s say you want to increase the number of opt-ins from your site. There are lots of different things you could test to make this happen.

You could try putting your opt-in form in a different location. You could change the wording of your opt-in offer. You could try offering a free gift instead of just a free newsletter.

But if you made all these changes at ONCE and your opt-in numbers began to skyrocket, how would you know which change was responsible? Besides, one of these changes could actually be having a negative impact — and you’d never even realize it!

If you have no way of knowing, then the next time you change your opt-in offer, you’re back to square one.

2. The MOST important numbers you need to know

You need to know TWO main numbers to keep on top of your website performance: the number of visitors you get and the number of sales you make in any given period. That way, you can figure out exactly how many visitors you need to attract to your site in order to make a sale.

Here’s how you do it:

Let’s say you got 300 visitors to your site in one day and you made 12 sales. Simply divide the number of visitors by the number of sales like this: 300/12 = 25.

For every 25 visitors you got that day, you made 1 sale. (Expressed as a percentage, this means your conversion rate was 4%, as 1/25 = 0.04)

If that conversion rate remains steady, then you can expect to make one sale for every 25 visitors. If you get 100 visitors, you’ll make 4 sales. If you get 1000 visitors, you should make 40 sales.

It’s a simple conversion — something everyone should know. It tells you EXACTLY how well your website is doing its job.

If that conversion rate drops for some reason, it could mean there’s a problem with your website — and you should do a thorough examination of it right away.

On the other hand, if it spikes, then you should try to figure out why more visitors are suddenly buying your products — so you can repeat this success!

3. Test the most important things FIRST

Don’t start off trying to track 30 things at once! Focus on your money makers first, as they’ll have the greatest impact on your business.

If your site sells more than one product, then you should be tracking the conversion rate for each product separately. Maybe your site is doing a great job at selling one product, but a lousy job at selling another. This is something you need to know, so you can figure out what’s going right in the first case and what’s going wrong in the second.

4. Think in terms of RATIOS as opposed to ABSOLUTES

Think in terms of percentages, not whole numbers.

Instead of thinking, "Hey, I made 25 sales last week!" — think, "Hey, I made 25 sales per 1000 visitors last week!"

Knowing you made 25 sales is nice — but it doesn’t tell you anything by itself.

If you know you made 25 sales per 1000 visitors, however, you know your conversion rate is 2.5%. (In other words, 2.5% of all your visitors last week bought something from you.) If it was a normal week, then going forward you can expect that 2.5% of all your customers will probably buy something.

This kind of knowledge is ESSENTIAL if you want to forecast how much revenue you’re going to make in any given time period. It also gives you a baseline number that you can refer back to when testing different elements of your web site in order to grow your income bigger.

For example: imagine if you changed the headline of your sales page this morning — and you ended up making 25 sales, just like you did yesterday. Does that mean that the change to your headline had no effect on your sales numbers?

Not necessarily.

Maybe you only got 100 visitors to your site today, instead of 1000. That means your conversion rate today is actually 25% — which is HUGE! That means 1 in every 4 people is taking action on your site. Wow. That must have been some headline you wrote.

(However, you really should find out why you only got 100 visitors to your site today. Once you bring that number back up to 1000, then if your conversion rate stays the same, you’ll be making 250 sales a day!)

5. Track your conversions by SOURCE

It’s not just enough to know what your conversion rate is. You’ve got to know where your best-converting visitors are coming from.

For example, are your most qualified visitors coming from pay-per-click ads you’re running through Yahoo Search Marketing? Or are they coming from your organic listing on Google — or maybe your affiliate network?

This is something you need to know! It’ll help you understand where your most lucrative stream of traffic is coming from — so you can focus your efforts there and work on growing those high-converting traffic numbers even bigger.

I’m interested in hearing how your testing is performing… please leave me a comment.