Data Informed Web Design

Posted on 03. Aug, 2012 by in Digital Marketing, Stats & Trends



Data informed web design is the backbone of our conversion optimisation process used to create on-going iterative improvements for our clients. It can also be used to de-risk redesigns, fix specific performance issues and inform design or strategy decisions.

Data Informed v Data Driven

Data driven game design has been around for years. Most top rate web and app designers you know will have been doing flavours of data driven design for years, but a design processes that incorporates data is not just in our past, but also our present and our future.

A redesign is seen by many companies as a high risk activity. What if it doesn’t work? What if it’s not accepted by users? Data driven design, in addition to reducing complexity & prototyping, is used to de-risk the design process. However, using data purely to drive the creative process also has its dangers when the data appears to go against business strategy or impacts brand identity.

An Informed Design Process

To mitigate the risks of digital design work, but still allow for creativity and intuition – those wonderful factors that breed innovation, we use data to inform rather than drive our web design process:

  1. Establish your data sources: Use a combination of data sources to build up as accurate a picture as possible of what is happening and why. Qualitative tools like web analytics and testing tools can tell us what has happened, qualitative forms of research like user testing and field research can give us insight into how people feel and think. Together these can be used to inform decisions and funnel into higher level strategy.
  2. Have a clear goal: Analytics data can be used to help understand how people are using the system, how they engage and interact with content and can be used to optimise conversion funnels and key user journeys. Facebook are also looking at ways in which they can use their data sources to quantify brand perception and long term network value.

    How you interpret and use data is important. To effectively model and analyse the data to gain clear insights we first need a clear idea of what we’re trying to measure and why. This must be linked to business goals and strategy.

  3. Build a micro and macro picture: Analytics data can help you focus on resolving specific issues. Designing for one element rather than the whole experience could be detrimental to overall site performance. When fixing individual design problems we always work to ensure we’re aware of the impact on other sections of the site and media.

    A good example of this is Harley Medical. Their analytics data told us that a specific page was instrumental in converting visitors into leads and was a common page within other top converting user paths. What wasn’t clear from the data was whether users then got in touch because the page content supported their decision making process, or didn’t provide the answers – triggering them to get in touch for specific info.

    We used our knowledge of user feelings/thought processes to create two alternative variations of the page. One that better answered their questions about finance (A) and another that made it clear they need to get in touch for their answers (B). When tested alongside the control copy version B significantly outperformed the control content and version A by almost 14%. Looking back at the web analytics we could also see that variation B continued to contribute positively in most top converting user paths.

  4. Use your intuition: How many times as a designer have you seen a dip in performance following a design change only to see performance improve over time and eventually exceed your previous benchmark?

    A combination of data sources allows us to build up an accurate picture of what is happening and why but this shouldn’t be the only basis for our design decisions. We must also incorporate other sources such as business strategy and regulatory requirements. Above all as designers we need to use our intuition.

    Data should guide us but not lead us. It is a resource we exploit to support our intuitions, helping us drive forward innovative designs and performance improvements.

  5. Evaluate your design: Any new design must answer a specific need whether it is to increase conversion rates, improve user engagement or raise brand awareness and extend brand reach.

    By taking a performance benchmark prior to the design change, using data to advocate change and measuring performance post implementation we can use these metrics to evaluate the design against the brief and showcase performance improvements.

    While it is nearly impossible to predict or forecast any uplift in performance ahead of implementing design changes, these measurements can help us clearly show how we make good decisions for our clients based on their specific needs.

Design decisions that have some basis in fact are less likely to fail catastrophically. By using data to inform the design process we’re not blindly seeking something that works better at engaging users. Instead, we take what is known and can be measured to become better, smarter designers.

Tags: , , , , , , , , , , , ,