Our journey to get closer to our customers by re-inventing how websites are put together.
Websites are frequently frustrating and annoying, particularly the websites of large corporations. Despite being administered by an army of experts in layout, design, analytics and optimization a simple rule mostly applies: user satisfaction is an inverse function of the number of pages on the site.
This is a problem if you have literally millions of pages to manage and it comes about due to the impossibility of creating a journey that works for millions of different users that have millions of different needs.
So, we had a simple thought: What if instead of expecting the user to navigate our website, we had our website navigate our users.
More specifically: Today users need to find the content that is relevant for them by finding their way through the site using information and signposting that we set out in advance. We can get that right most of the time, but it is hard to get that right in every single possible use case.
Our new approach, called Fluid Web, uses data to get an understanding of the likely needs of each user and then intelligently selects relevant content using more data together with AI driven algorithms. As the user navigates the site, we improve our understanding of their needs and continually improve the experience.
Ultimately, we learn enough that we can fully anticipate the user journey and provide a (near) perfect experience.
Would that really work in practice?
Spoiler alert: first indications show that it does actually work.
What started out as a “what if” idea turned in to a project involving 27 experts from 10 countries over a period of two years as we refined the idea, understood the challenges, developed the technologies and built and tested prototypes.
Simulations based on historical web data showed us that we could reduce the number of clicks in a typical user journey by up to 30% – very roughly translating in to a 30% decrease in irritation, or a 30% increase in positive results.
You can try it today
On the1st of October this year, we released the first public version of Fluid Web. Normally, we would do this on a set of low profile, low traffic pages so that we could quietly iterate. However, AI is all about data and so volume is key. And when we wanted to redesign the front page of siemens.com to support our new brand design we wanted to make it more than just a visual design, but a truly intelligent experience.
Therefore, version 1.0 of Fluid Web is now powering the front page of siemens.com, meaning that certain parts of the navigation and the content you see there has been intelligently selected, personally, just for you.
We are still in the very early stages of refining and perfecting this approach (don’t be surprised if it doesn’t quite feel like it was designed just for you) but early data shows that we are around 200% more effective in getting our users to the information that they are looking for and that we are significantly reducing the number of pages shown that users are not looking for.
First steps, next steps
So far, we have shown that the approach works theoretically, and now we need to prove that it works at scale – which is a different challenge.
Over the coming months we are going to apply Fluid web to additional parts of our web presence to get additional feedback and optimize more of the journey.
More importantly, our developmental journey has given us a lot of insight into the challenges of working with new technologies that have the potential to completely up-end the way we see existing problems.
The future of Fluid web is going to be very exciting, but the future of the ideas and possibilities that the underlying technology has created will be more exiting still.