The vast majority of websites are designed in the hopes that visitors will complete an action. Buying, subscribing, commenting… from massive social media juggernauts like Facebook, to small artisan websites selling local, hand-made crafts, this is true. Marketing has always been about understanding the subtle ways which presentation, repetition, and other visual or sensory artifacts can influence the decision-making patterns of consumers. Marketing, almost since it’s inception, has relied on data – in the form of research of all kinds – to achieve its end goal. So, data-driven web design is, all-told, simply any kind of website design which utilized visitor data to modify design for optimal results.
The ability of websites to respond to data has dramatically improved in recent years. The advent of easily-parseable analytics platforms, self-modifying websites, and other fantastic tech tools have allowed website designers to modify their product to more optimally meet the goals of the owner in real-time. We’re going to talk today about using data to draw the line between visitor data and design improvements which maximize benefit both to the owner and the end user.
Because it is also true that web design should always serve the visitor, a design which is overcomplicated or tedious for users will almost always fail, in some respect, in the goal of its use.
Sure, marketing or user data has almost always been used to influence and guide web design to some degree, but what does data-driven design mean for us today? The answer to that question requires taking a good look at what data is. Most data is either quantitative (numerical) or qualitative (non-numerical). And the vast majority of easily-accessible data is quantitative. For example, Google Analytics informing you that 10 visitors viewed one blog page, while 1 visitor stopped by another.
Qualitative data is important. Following the numbers can help you find what visitors value and what they do. But it doesn’t tell you why they like it. And understanding that requires qualitative data. The best web design companies are pushing the field by getting qualitative data and modifying their designs based on this data. They’re using beta-testing systems and getting powerful, direct feedback.
A good data-driven design program will always use both kinds of data. Testing, analysis, and feedback systems are continuous, rather than one-off, and therefore provide deeper insights on what visitors prefer and want the websites they visit to do for them. It is this kind of deep-digging which pushes the horizon of design forward in ways which are meaningful for users.
Getting Specific, Actionable Data:
Whether the data used in data-driven design is numerical or no, it must always be empirical, and address specific questions. Whether you’re a UX designer or a data analyst, a little bit of scientific inquiry comes into play at some point, because you’re not just wanting to look at aggregate data… but to answer specific questions.
Hypothetically, this could look like a situation in which a blog or online magazine, which has hired a designer to update their website. The goal of the website is to keep visitors engaged and to provide useful information, but also to increase subscriptions. As a designer, those are several goals to meet, and ideally you should begin studying the existing website to identify key areas of improvement.
Key quantitative data metrics you might study include the bounce and exit rates, and find out which pages have the highest in each category. This provides information on a ‘what’ which you should try to address! From there, you might use UX testing to get qualitative data from visitors on their experiences from these pages to find out the why. User-reported information might reveal that page-specific CSS makes exit links larger or more obvious on these pages, therefore funneling users away. Lots of moving elements might bog down the pages, providing a negative user experience.
But in this hypothetical situation, you have used empirical information addressing specific issues to yield information which should influence design decisions. And when you start employing this one-two punch to achieve great design, you’ll realize some amazing trends.
The Big Takeaways of Data-Driven Design:
Applying data-driven design has helped push design elements which were once novel and boutique into the realm of commonality. Responsiveness, high-contrast CTAs, flat design elements were all once thoroughly tested and checked by data-driven design teams. But that’s just the beginning! Increasingly, data-driven design is finding that design decisions should be tailored to specific audiences. Right now, this is most commonly seen as different landing pages or page elements for different geolocated areas. But designers are even experimenting with segmenting design by audiences even more: by age, for example.
Testing for data and then using that data to modify design can be resource-intensive, that’s true; but even small, focused efforts can result in big improvements for clients, and the increasing ability to meet the goals of the websites we’re building.