Web Testing

What is A/B testing? It is a process where web pages are compared with one another. Typically the ones which are compared are the landing pages as this is the most important page on your site. A/B testing compares several alternative versions of the same web page. An example of A/B testing is first decide on the page or application you want to test, such as trying to see how many users register on the landing page. Then you want to see if you can improve the percentage of people who register by altering the landing page. A/B testing can only test one baseline version against the new version simultaneously.

With all statistical analyses the quality of the data is improved with an increase in the size of the number of users and the length of time the testing occurs. So the rule is to set the time as long as possible when carrying out A/B testing. Implementations are difficult to analyse in the short term, as chance is often a part of the results.

Targets need to be set for a successful exercise in testing conversion rates. This testing can be for average order level/size, revenue per day, time spent on the site etc. Performance goals have a very important role in an e-commerce site as they can represent the success and return on capital for the owner.

Websites often need to personalise the user experience. Clearly, this would be impossible to achieve for every user but small changes often have dramatic effects such as when a user goes to your landing page, the banner reflects the fact that the site is able to supply its goods in your country. An e-commerce site needs to capture the IP address of each visitor to your web site and then relate the IP address to a geographical location. You can set content-specific information on the site so geographically the site reflects the country of the user accessing the site. If you supply goods and service around the world then informing your customers of that fact will achieve higher sales. Try and ensure that the customer is aware of your supply criteria.

With Google AdWords you need to make sure the advert reflects the target web page. The continuity is increased if the AdWords reference is targeted against the corresponding web page. Remember a user will only stay on your site if they feel they will achieve their goal. Any distractions are a waste of time and will send users away. Internet shoppers are easily distracted and even when they have an intention to purchase the users can often slip away before making a purchase. To enable the echoing of AdWords content is straight forward as you need to marry up the URL for the page against the AdWords advert. This linking of pages to ads holds true for affiliates and email marketing.

When shopping online, the user tends to be acutely aware of the need to check the price, shipping costs and the time of the delivery. Customers often require help in choosing the design or service. So on your web site you can write a “badge” which shows e.g. this is “universal” or “this is our best seller” or “lowest price”, each badge or label will give the customer that added security that they are making the right decision. Distance selling needs to show a high level of trust. Badgeing or labelling will make the customer realise that you have more knowledge of the product compared to your competitors. The best way to show the badge is to show them as overlays. This will allow you to compare the effectiveness of each overlay so for example you can start with “hot pick” and if that is not effective then this can be changed to “staff pick”.

Following on from A/B testing is the variant of MVT or Multi Variable Testing which is a more complex version of A/B testing. Users hate leaving email addresses when going to the checkout because they believe their email address is going to be used by spammers or sold to marketing companies. If a small explanation is placed next to the request for email stating that the email will not be used for any other purpose other than as a confirmation of a purchase by the user then the user will be more confident in providing an email address. Customer experience has a very fragile hold on your website and any tools to keep it in place are very important.

MVT and A/B testing give quantitative information on the impact of changes made to your website. These results start from the site being baseline at say 100 and then the changes can be evaluated to see whether the changes have increased the score from 100. With MVT Testing you can use up to 9 different variants for testing and by scoring the success and failures you can see were changes can improve the success of the site.

The crucial question is how do you persuade users to spend money on your site. You will spend a fortune attracting possible customers through AdWords or digital marketing and the user will then see your site and will know straight away whether to stay or go. Focusing on SEO, SEM advertising and email marketing is only one side of the picture, you have to convert the possible customer into a sale, hence the need for an online marketing team. This team has to work together as individual efforts will reduce the effectiveness of the team. Collaboration working is efficient and cost effective therefore data produced from the websites need to be broadcast to the other team members for analysis. This information can often be relevant to other aspects of your digital marketing campaign. The stages of the buying process need to be understood so decisions can be made with regard to the digital marketing approach. When does the user “buy”, where does the user live, does the user buy on impulse or from necessity. These questions are familiar to marketing an approach in the high street shops and the same is true of how users buy goods and services on the internet. Once you are aware of your customer’s interests and habits you can connect better by using this information in the digital marketing plan. New ideas need to be thoroughly tested before they are put into the mainstream. Regular customers may find the changes unsatisfactory and will move away to a competitor’s site. All the value of those regular customers will have been lost. Any changes to the website are extremely important and therefore have to be carefully tested.

One of the best ways of testing your website is to use Google Web Optimiser which is free. This tool will give you a start and an understanding of the changes that you are making to your website and their effects. Google optimiser is for websites that have less than 10,000 hits per month. E-commerce sites which have higher hits really need to buy a more accurate tool for analysis.

The customer is king and you need to ensure the e-commerce site moves the way your customers want to move. Websites that make users jump through hoops before getting what they want are not a very good idea.

So remember the customer is king!