How To Increase Your Conversion Rate
Everybody talks about the importance of testing your sales copy or a page layout. After all, proper testing can help you modify your page in a way that will drastically increase your conversion rate.
In this article, I would like to describe a way to shorten the amount of time it takes to test your pages and to increase the probability of success.
Before I go any further, I would like to introduce a few concepts and notions that will be used in this article.
Attribute — a specific visual or conceptual element of a page, an ad creative, or a sales letter (used in fine-grained performance comparison testing). A few examples of what might be considered attributes: – headline text – headline font – headline color – order button size – order button color – order button text
Please note that even though those six things are related to only two elements, they are all separate attributes.
Attribute value – some particular setting of an attribute. Here are a few examples of values: – order button color, red – order button color, green – order button text, “Buy Now” – order button text, “Add To Cart”
I just listed two values for two separate attributes.
Significant attribute — an attribute that affects the performance of a page.
Insignificant attribute – an attribute that does not affect (or has little effect on) the performance of a page.
There are some obvious significant attributes that are universal for everybody.
One example of such significant attribute is a headline. It has been proven many times over that changes in a headline have a huge impact on the performance of a campaign or an offering, in any medium for any industry. You can find a lot of information about universal significant attributes in any book that deals with testing and response rates.
A much harder problem would be trying to identify significant attributes that are unique to your site, product, audience, or traffic source.
As I described in my report called “How To Win The AdWords Game,” the famous 20/80 rule applies to attribute testing just as well as it applies to many other things in our lives. In other words, 20% of the attributes you improve will produce 80% of overall performance increase.
Out of 100 attributes you decide to test, testing 80 attributes would be a waste of time. This is the reason many people fail to realize the importance of small attributes.
After all, if you follow the conventional wisdom of testing only one attribute at a time, you end up with no visible results and a firm belief that small attributes do not affect conversion. It is only logical to quit after testing 10 different attributes, one at a time, and having to wait one week for each attribute. The truth is, you have most likely spent that 10 weeks testing your insignificant attributes.
Since there is no way to know in advance which ones of your attributes are significant, the only reasonable thing to do is to test. You need to test and find out which attributes have the most effect on your visitors’ behavior before you start testing different values of those attributes.
Let me give you a simple example of what I mean: You need to establish that a color of an order button is in fact a significant attribute before attempting to find the best producing color for that button. If you start testing different colors when that attribute is not significant, you just waste your time.
So how can you find which attributes are significant and which are not in a reasonable amount of time? It’s simple. You need to test in parallel.
You need to think up as many different attributes as you can and create different values for each of them. After that, you need to present a random set of attribute values to each new visitor, and keep the same values for returning visitors. Once you do that, you need to collect and track your test data to measure performance based on the sets of values.
For example, let’s assume you tried the following attributes (with a set of values): – a color of an order button: blue, green – a text of an order button: “Buy Now”, “Add To Cart” – a color of the font that lists the price: red, black
That way, one visitor might see a blue “Buy Now” button next to the red price, while another one might see a green “Add To Cart” button with the black price, and yet another one might see a green “Buy Now” button with the red price, and so on.
With this set-up, you get 8 combinations of three attributes.
Once you ran a test, you got the following conversion rates: order button, blue = 1.53% order button, green = 1.52%
text of a button, “Buy Now” = 1.95% text of a button, “Add To Cart” = 1.01%
color of price, red = 1.51% color of price, black = 1.49%
From this data, you can tell that you got the most performance difference by changing the text of an order button. This is your significant attribute. Forget about the other two for now and start testing the text of a button.
You can take this concept a step further and test combinations of attributes. You might find that changing a color of the price together with a text of the order button produces better results than changing the color alone. I will not cover this topic now, but will write about it in the near future.
For now, let’s just concentrate on picking stand-alone attributes that show to be significant to the performance of your page.
Once you have identified those attributes, it’s time to start tweaking their values and test results, also in parallel. You need to apply the same concept, to testing values of attributes this time.
Keep in mind, that small attributes are often unique to your site and your audience. What might work for you, might not work for other people. Nevertheless, if you can correctly identify your small, but significant attributes, you should be able to increase your conversion rate. The effect of those small attributes might not be as significant as with headlines or other well-known attributes, but the more attributes you find and optimize the higher you increase the overall performance of your page.
Konstantin Goudkov studies the psychology of pricing, ways
to manipulate prices for maximum profit, and tactics to
control consumer price perception.
You can find his latest report at: