Conversion optimization and A/B testing

Conversion optimization and A/B testing

Optimizing the conversions of your website/blog means making money. If you don't know where to start, we'll help you with this little guide. A reliable and effective method to achieve conversion optimization is to carry out “split testing” or “A/B testing”.

What is conversion optimization?

In poche parole optimization della conversione significa aumentare il numero di visitatori del tuo sito web/blog che effettuano “realmente” un acquisto finale. Ma sono anche quei visitatori che incoraggiano uno specifico comportamento degli altri visitatori, come per esempio iscriversi ad una newsletter o partecipare ad un report gratuito.

To try you have to try! What does it mean? You will never know how your website/blog will react if you don't make changes. Once you have done and changed individual elements you will need to measure the impact of these changes on visitors' actions. The first few times you won't know what to do, but over time you will be able to perfect your website/blog.

The advantages

Non è vero che più aumenta il traffico verso il proprio sito web/blog e più si guadagna. Devi concentrarti molto sul contenuto dei tuoi articoli, sulla SEO, sulla pubblicazione sui social media, la pubblicità e il nuovo contenuto.

Focusing on advertising and marketing is certainly a good basis for success, but don't forget your real visitors, they may feel neglected and you could miss out on a lot of revenue.

Correct conversion optimization, which must first be understood and increased, can make you achieve much more sales and earnings with the same number of visitors. Often a very small change or an imperceptible modification causes your website/blog to have many positive effects in return.

In this article you will also find many ideas on how to improve your website/blog.

Conversion optimization

If you are wondering whether you should focus only on conversion optimization we recommend you try "split testing" or "A/B testing" with your visitors.

Practical example 1:

you have 100 visitors per month on your website, but it may be difficult to obtain real statistical results due to the sample being too small. The result of the A/B test is: 50 visitors see button A and 50 see button B. These figures are not a sufficient quantitative measure, but you can always try to obtain useful information.

Practical example 2:

You have more than a thousand visitors, then you should definitely implement conversion optimization. To do this you can use plugins that are very simple to integrate and can be connected and configured with just a few clicks.

A/B tests

How does an A/B test work? It's very simple. First you need to create two versions of a site element and test them to see which one works best. You could try a different style, an order button type, or maybe a sales page with and without videos.

Always remember to make only one change. If, on the other hand, you have revolutionized too many elements you will never know which of your changes was the right one, or the wrong one to avoid.

We recommend that you do the parallel comparison. The results you get must be compared with the same time, the same day and with the same methods.

Optimizely.com

If you find it easy to do A/B testing on your own, you can do A/B testing without external tools. The easiest method is to create two product sales pages and create a different element for each page. Complete the page with the alternating configuration of the two pages to be shown to visitors alternately so that they can view both product pages.

When you have decided which period to take into consideration, you will need to measure the results with a statistical tool, to compare the clicks and conversion rate of the two versions. Check before doing everything because it could be expensive, cost you a lot of time and be complicated.

However, if you think that carrying out A/B testing is a little difficult for you, there are external programs that can help you. For example you can implement with certain conversion optimization tools, such as optimizely.com.

You can register online and carry out an experiment for free. The procedure is simple:

log in

click on “new experiment”

give a name and enter the URL of the optimized page

press enter and you will have the result.

Immediately afterwards create a modification, and change an element of the page, which you can do directly in the optimizely.com tool, without having to change the code of the website/blog. Thanks to this tool you can make various changes and see what happens. We advise you to limit yourself to just two variations.

The next move is to specify a goal regarding user behavior, for example if the user finalizes the purchase, you could focus on a “Thank you for purchasing” store page; or a newsletter registration confirmation page, etc.

With Optimizely though, you can't enter the tracking code if you want to improve your affiliate earnings. In fact, you cannot specify the partner program store site as the final destination, but you can count clicks on an affiliate link.

Lastly you will have to be patient. The data is not collected in a few seconds, but you have to wait until there is enough data to evaluate which option works best. When the work is finished you will know which version is the best, and then dive headfirst into your website/blog with this version.

Remember to do one A/B test at a time, first finish one test and then you can start another. Furthermore, when the test is running you must embed a reference to the Optimizely cookie in the privacy clauses of your website/blog and provide an opt-out link.

The relevance

How to decide which variations to modify. It would seem like an easy choice, but it is not and often two variants lead to very similar results, in this case the similarity is not reliability. In statistical terms, “relevance” reliability is measured with numbers from 0 to 100 as a percentage. Obviously the higher the percentage, the greater the relevance and therefore the reliability. But if for example the result was 90%, it means that a 10% offers the possibility that the two results are the same. Always try to reach at least 95% of relevance and you can also use Optimizely.com as an analysis tool that allows you to identify relevance in the results of each variant.