E-commerce web design based on user tests

Every web designer and e-commerce customer knows what a good e-commerce needs to have: a shopping basket, buttons for adding items, products lists organised by categories and a search engine. A payment process will also need to be set up, which may or may not require prior registration and, once the shopping process is finished, an automated email sent to customers with the shopping details.  

Many websites make what should be simple more complicated , and sometimes their innovation is not focused on being more useful or, even worse, they do not accept that the standard model could be valid for their product line or for their customers.

Reality tells us that, if there were answers to every question people made about e-commerce, small companies would not have any space on the Internet, but this is not the case, and SMEs capture a larger share of the Internet market than might be expected.  We all have a premise in mind: we all would like conversions to sales to be as high as possible.

There are no shortage of people with an in-depth knowledge of purchasing behaviours,  particularly of  online channels and also of the relationship between design and marketing. Although we do have access to this knowledge,  no-one has a magical formula that always works.

It is important to bear in mind that selling cakes, bathroom furniture or holidays are not the same thing (whether or not they are perishable products is not the only difference). The web management needs are not the same, the people you will need to deal with are going to be different and also the customer’s expectations. Every company needs a bespoke e-commerce solution.

There are many answers based on research, experience…but most decisions about design are based on personal preferences or expectations rather than facts. This is why at Imaginanet we decided to integrate experimenting with solutions into our content management system.

Now, let’s imagine we have just started an e-commerce business.

Some typical questions would be:

Which promotions have generated more income: 2 for 1 promo, 20% discount or free delivery?

Which colour combinations help to sell most products?

How many links to related products should there be on every product card? Would 3 or 10 be tetter?

Does using large images produce better sales results than medium ones?

The answer to most of these question will be “It depends”. It depends on our customer’s profile, on the product or service we are offering, and on many other factors. There is not a definite answer to each question.

Not very long ago, e-commerce sites tended to imitate Amazon on the basis that “if it works for Amazon, it will also work for us”, but buying books is not the same as booking your holiday or looking for new furniture for your house. Increasingly, customers are looking for virtual user-experiences that are more more closely aligned with the product or service they are looking for, and are less willing to tolerate a mediocre level of usability and design. In addition, we need to bear in mind that the customer does not have any particular sense of obligation or commitment to our e-commerce site.

All this has led us to focus on researching  and developing tools that make it possible to experiment. In this way, we can provide two different interfaces for a website (one with more adverts, and the other with fewer, for example), and we can present them to users in a random way and thus measure the popularity and effectiveness of each using values such as length of visit, sales completed, most popular orders, etc.

Creating two different interfaces is not something new, as many web design companies are offering it already. Our innovation lies in being able to experiment with the same tools that are being used daily on the website. All the data relates to benefits, which at the end of the test period will allow us to know with certainty whether the colour green works better than white for sales, for example.

For us, experimentation is a technique to be applied which enables us to enter into a dynamic of continuous improvement instead of that of annual re-design, where you could risk discarding precisely those design features which were actually working well.

We can also identify the exact point in the purchasing or payment process where the user moves backwards or withdraws, as well as assessing the most common error points.  Rather than looking for these less efficient areas of the web, we can target these for you, enabling you to focus on re-designing them or improving  the content.  

Experimenting with design and usability translates to being able to make better decisions, based on customers instead of personal preferences. Questions about what product descriptions should be like or whether the payment process should be done in 3 or 7 pages are not immovable. What could be better than being able to choose the method, design and performance that works best for your  particular clients, and if doing tests is easy, this allows time to make changes.

In our opinion, if  the decisions concerning  design and functionality are not based on experimentation and market research, the results will never be exceptional. In those cases where problems with the design are suspected, or where it could be better or more efficient, systematic  testing is needed  so that the results can provide more accurate confirmation of what the problem is and where the solutions lies.

Web design, usability and natural positioning within search engines

We combine attractive web design for users, simple web browsing, where customers can easily find what they are looking for, with coherence with the company brand image.