Versio2 welcomes you to its latest venture, restaurant Le Marketeur Petite, a simple introduction to data-driven site design. From A/B testing to eye tracking, data drives intelligent website design, and you need it in your life.
What is data-driven design
Imagine youre the new owner of a fancy French restaurant, lets call it Le Marketeur Petite. Why not. Its popular with locals, but there’s always room for improvement. Your customers come in, they sit down and read their basic menus essentially a services list with pricing and then order i.e. convert.
Now, lets apply data-driven design principles to Le Marketeur Petite. You add thumbnail images to your menus and find that those dishes pictured become the most popular. What next? Print two different menus, with different images, and give one variant to one group of tables, and the other to another group what do you find? Then you make a note of how young and old customers order, or perhaps find that people in your area prefer particular types of food compared to those from other neighbourhoods. One day you read a study on how people read pages, so you try different menu variants with special offer call-to-actions (CTAs) at the bottom of your pages click here for a great article on CTA design.
After some weeks, youll have gathered lots of lovely data. You perform proper analysis and learn which dishes (services) perform better for your target customer; where to place images and special offer CTAs; and how to make particular dishes the higher-margin ones more desirable.
The principles at work
In our restaurant analogy we employed some basic data-driven design principles that are as applicable to designing and optimising a menu as they are to a website. In fact, analytics tools make it even easier:
- Use site overlay analytics Google Analytics, just one example of a popular traffic analysis tool, has a site overlay feature that shows how many users click on each of your website links as a percentage of overall clicks. Try it out and youll soon start to see important patterns is one CTA doing particularly well, and if so, why?
- A/B test as standard Also known as split testing, this is the practice of creating two or more variants of a page with different images, text and/or CTAs, to see which garners more conversions. Its good practice to split test landing pages for special offers, especially if visitors arrive there from other marketing channels such as Facebook.
- Don’t be scared by ethnography Not quite as sinister as it sounds, this is simply exploring user behaviour in the context of cultural phenomena. Some analytics tools and online marketing platforms can provide user demographics obvious ones are language (from browser settings), and region (from IP address mostly) and then provide metrics to show how different users behave. Which demographics are you targeting? Find out what works with them, and focus your efforts there i.e. optimise.
- Read eye-tracking studies Theres a wealth of data out there on how people read things, dating as far back as the 1800s. Studies of human-computer interaction came in the 1980s, and along the way we picked up the now-familiar Z-layout concept of web design. Have you ever wondered why websites generally have their logos in the top-left corner? This image from Webdesigntuts+ demonstrates basic Z-layout theory.
The case for data-driven design
A Forrester Consulting report from April this year a survey of 209 marketer, eBusiness, and customer experience professionals found that 60% have seen website improvements due to the use of data. Further, 75% say companies are looking for new ways to optimise their website(s) through the use of data; and 57% agree that multiple data sources are used when designing and optimising.
If youre new to data-driven design and dont believe its quite your thing, have you ever argued with a designer over where your logo should go, or where to place special offers? Raw data holds the answers.