Custom-made User Experiences

In order to test changes on the user interface, improvements to algorithms, further development in apps or content management systems, thousands of experiments are carried out every year on the websites of large companies such as Amazon, Bing, Facebook, Google, LinkedIn and Yahoo. "Our success depends on how many experiments we conduct per year, per month, per week, per day," said Amazon CEO Jeff Bezos while explaining the strategy at the end of 2017. Online-controlled experiments are now considered an indispensable tool – not only for large companies, but also for startups and smaller websites.

At the same time, the collection and evaluation of customer data is gaining exponentially in importance. Creating user profiles through web tracking, linking web, CRM, and other data, big data records and automating customer communications are just a few pieces of the puzzle on the road towards a new paradigm: instead of the "One size fits all" solutions used at the beginning of website development, the individual customer approach is used. Today, the optimal user experience is a custom-made product for each individual customer, which accompanies them along their entire customer journey.

To facilitate the creation of a perfect user experience, high-quality software solutions are now available that specialise in quickly evaluating ideas with the help of controlled experiments - also called A/B tests, split tests, randomised experiments, control/treatment tests and online field experiments.

However, many companies find it difficult to integrate the experimental processes into their workflows and to take concrete steps in carrying out experiments, such as defining suitable test scenarios. This is where Portaltech Reply comes into play with a comprehensive range of consulting services right up to enterprise solutions.

This is how it works

As a form of preparation – especially in larger companies – an explorative inventory of the current website or product development process, for example via expert interviews, can be useful in order to formulate clear goals, responsibilities and a roadmap for the introduction phase based on this. The meaningful integration of new tasks into existing workflows can also be planned ahead of time and possible reservations of those involved can be identified early on.

A correctly implemented experimentation culture can quickly lead to a democratisation of idea generation for optimisation processes: it replaces gut decisions or the general implementation of the highest paid person's ideas. Often, these advantages become apparent early-on during the training phase, for which Portaltech Reply offers various courses, including training on the most appropriate experiment strategy (how and when should which experiments be carried out?) or experiment implementation (tool configuration, QA and technical implementation).

A critical point in the implementation is the finding of meaningful optimisation hypotheses. The experts from Portaltech Reply can provide assistance with this as well. In the methods they use, quantitative and qualitative data are linked to form an overall picture. The former can be collected through web analytics, heat maps, A/B testing, BI data and market research. Qualitative data, on the other hand, can be generated through expert evaluations, usability labs, remote tests, surveys and customer service input.

Isn't this a waste of time?

At first, testing and experimenting sounds like a very time-consuming process, and time is a resource that is only reluctantly used generously. However, with the right tools and an experienced partner like Portaltech Reply at your side, you can experiment and optimise an online shop without burdening your IT resources with unnecessary, open-ended projects.

In order to offer its customers a successful experiment setup, Portaltech Reply has entered into a strategic partnership with industry leader Optimizely, so as to support one of the tools that helps marketing and product teams test ideas, gain insight and create outstanding user experiences.

43 percent more turnover

The figures that have already been obtained from Optimizely users in various use cases are impressive. One example relates to the Jawbone fitness wristbands. Before the company based the development of its shop on a valid data base and confirmed hypotheses to increase its turnover, Jawbone's homepage showed the various fitness trackers of the UP3 brand side by side. After testing a few variations, the company created a version of the site that did not focus on the individual products, but instead on the presentation of the benefits of a fitness wristband. The site focused on highlighting the advantages of the fitness gadgets to the users, in order to encourage them to buy one. A strategy that proved successful: turnover per page visit rose 43 percent in the desktop version and 24 percent in the mobile version.

An effective strategy was also adopted by Electronic Arts for the market launch of a new edition of its computer game Simcity. On the product page that was initially created, a price reduction for pre-ordering was advertised in a prominent place. However, the version of the site that managed to increase sales by 43.4 percent completely abandoned the pre-order offer, and instead merely presented just two purchase options for the game. Whereas pure product presentation meant lower sales for Jawbone, this type of webshop design proved to be more suitable for the online gaming world's target audience.

Turning in circles – innovatively

However, one thing is certain, data-driven experimentation rarely means stopping at the next best variant of an online shop; rather, experimentation should be understood as an ongoing process, as a cycle that repeatedly goes through the steps of brainstorming, planning, development, analysis and implementation. That’s because new hypotheses should always include individual tests and campaigns. Once new tests have been developed and launched, they need to be analysed and decisions made, on the basis of the results, as to how further action should be taken. After a successful implementation, it should be questioned and improved again and again with new ideas and hypotheses.