CX & Design June 25, 2018
Performance is User Experience
Currently, user experience is on everyone’s lips. Especially in e-commerce, it is an essential topic as it directly links to the conversion rate. But the question what exactly defines good user experience is answered differently by different people: excellent design, intuitive user guidance, a wide variety of payment options or a well-elaborated self-service area.
Funnily enough, very few will spontaneously answer this question by saying: “To me, it is important that the shop is extremely fast at all touch points!” However, this is a very obvious answer for us. Just imagine a website taking ages to load when you want to shop online. Does this make you feel pleased? Of course not.
Sufficient speed is a basic requirement and, quite naturally, taken for granted when planning a shop. Unfortunately, when it comes to choosing the platform, we tend to forget to thoroughly check the mandatory prerequisites in advance. Too easily are we impressed by a plethora of special features and fancy add-ons. But it is difficult to mend mistakes in the foundation once the house has been built.
Google recently announced that page speed is to become a new ranking factor for mobile pages from July 2018 on — according to the fact that the user shall be provided with the best experience possible while looking for answers.
This is even more interesting for e-commerce companies: speaking of the correlation between speed and sales volume, that has been analysed in various studies. Thus, the Aberdeen Group has found out that the delay of only one second already leads to 11 % fewer page views, a 16 % reduced customer satisfaction and 7 % less conversion. Amazon mentions as well that each millisecond delay costs 1 % sales volume. Time is money — as Benjamin Franklin already noted in 1748.
Focus on speed often is considered when it is already too late.
Only when the shop has already been launched, the long feature backlog is completed, and no more major A/B tests are coming to one’s mind, it is often realised that the shop should actually be much faster. At this point, we as an agency are sometimes briefed by new clients to trim the shop to performance. But this can be compared to the cellar of a house — it is almost impossible to implement it belatedly with an already completed architecture.
The decision for performance and scalability is a fundamental one and therefore has to be integrated into the platform decision.
So far, only a few systems exist on the market that can solve the page speed issue with an appropriate architecture by themselves. These are the ones recommended by us, if we get the chance. However, the cooperation with a client often starts with an already existing shop based on a flawed architecture, and we are hired to subsequently modify it to achieve better results.
Sure, there are always workarounds to improve shops that are not fast enough due to their architecture. The most popular one is the widespread full-page caching platform Varnish. It can be closely connected to platforms such as OXID or Magento. However, it only compensates for the unsatisfactory software design, as online shops are very dynamic platforms (e.g. pricing, availability, personalised content). The aim of caching is to display dynamic data statically and thus reduce the response time of the server, since database queries are no longer necessary, for example. To do this, a dynamic page is created individually and then archived in the cache. This ensures that all ongoing visitors receive identical content. However, this creates additional complexities and challenges: When should pages be deleted from the cache? How do I communicate that a product is no longer available even though it is offered to the customer? How can I personalise my shop if I need to deliver identical content at all times?
Well designed e-commerce platforms provide the solution.
As a buyer of an e-commerce platform, you should expect that the respective manufacturers will provide solutions for these problems. Instead, they often rely on makeshift solutions and invest exclusively in the development of new features. The pitch still works this way, because many shop operators still buy their e-commerce platform concentrating on the length of the list of provided features rather than the quality of the system architecture. You should always check in detail to what extent the standard software really matches your own requirements, though. In many projects that I have accompanied until today, said extent was rarely particularly high. This is often due to the fact that shop operators and manufacturers have completely different views on how the function works “correctly”. Basically, you should ask yourself which special feature could improve the conversion rate as effectively as a short loading time.
In contrast, there are the online pure players, such as Amazon, Zalando or ABOUT YOU, that have identified good architecture as the key to success and, therefore, have developed their platforms themselves. And that pays off today: They are scalable, with extremely high-performance, highly personalised and able to work data-driven. This helps them to optimise the user experience in even more areas and to outmatch themselves time and again.
Spryker as a method of resolution
Spryker identified this problem and created their Commerce OS — a platform addressing the main problems of established shop systems and also being able to solve them. The focus is on a high-performance and scalable architecture that enables shop operators to compete despite the lack of huge development teams.
If you look at the average distribution of traffic on the individual pages of an online shop, you will soon find that at least 60 % of it is searching the catalogue in order to find the right information and products — with TV advertising the proportion is even higher.
Spryker have therefore converted a shop into two separate applications. The “customer-facing” frontend (Yves) and the backend with business logic (Zed). Both parts of the application can be scaled horizontally and separately from each other and can thus react to the particular individual needs of each customer. In contrast to conventional shop systems, the complete business logic is not executed for each individual access to the shop, but only when it is really relevant, e.g. during “add-to-cart”, login or checkout. As long as none of these actions is performed, all information is available in an extremely fast in-memory database, enabling delivery times of 50–70 msec.
Performance is one of the main factors for success in e-commerce today. Customers expect fast response times, whether it’s a sunny Wednesday morning or Black Friday. After all, the big players are demonstrating it to them and thus determine the standards of the market. That is why they are the ones that every online shop nowadays has to compete with. So, I can only recommend approaching the platform selection from a different perspective for the next relaunch: Don’t ask about feature XY, but rather about the foundation: How fast is the system? What do you do to avoid caching? Does the software architecture support performance-driven approach?
This is the only way to choose the platform with which your business will continue to scale over the next few years and with which you can react to the ever changing rules.