ui case studies

How & Why UI Case Studies Daniel Burka Pownce - Co-founder Digg - Creative Director Silverorange - Co-founder Case Studies 1. Digg - Story Format Getting started and changes over 3 years 2. Digg - Comment System Adaptation over 2.5 years 3. Pownce - Initial Design Starting from scratch Starting with Digg • • • Started working with Digg in March 2005 Original scope is one week of work Where do you start? Digg story design Why start so focused? • • • • • You need to start somewhere Homepage template was more obvious but... Identify key themes quickly Get fast & focused client (Kevin) feedback Develop visual language Example Mozilla Download Page & Button Example Store product page ? Identifying the problems ? Identify what makes Digg unique... Digging! 1. Doesn’t feel important 2. Doesn’t indicate an action 3. Not quickly scannable ? Title isn’t bad, but... 1. Could be emphasized better 2. Help user jump down page visually ? Submitter information 1. Oh, he’s the submitter! Not the author? 2. What the heck is that number? ? Report? 1. What do you mean? 2. Why would I do it? 3. That’s a multiple select dropdown? Really? ? More stuff to add? 1. New comments system 2. Ability to share stories 3. Show the source? 4. What order are these in? ! First revision... ! Ok, spot the differences... 1. Digg button is emphasized... a lot. 2. Digg button is an action with a verb 3. Button style helps with brand identity, visual language Mezzoblue (thanks Dave!!) Mozilla Digg ! What else has changed? 4. Title is bigger and bolder 5. Submitter is labeled as such and has an image 6. Time stamp is entirely new ! What else has changed? 7. Comments are entirely new 8. Two share links are new 9. ‘Report’ is now clearly a dropdown renamed ‘Bury Story’ Success? Success? • In this case not always clear cut • • Can people scan stories faster? How does it make people feel? Success? • • In this case not always clear cut • • • • Can people scan stories faster? How does it make people feel? More digging and other participation? More registered users? More overall visitors? Gather stats where possible Success? • • • In this case not always clear cut • • • • • • Can people scan stories faster? How does it make people feel? More digging and other participation? More registered users? More overall visitors? Is date-order more comprehensible? Can someone successfully digg or report? Gather stats where possible User test, even basics ? Identify the problems again... 1. Digg number is too big (the site’s getting popular) 2. Gradients are so last year 3. Too much stuff - could be tighter ? Identify the problems again... 4. Meta data divides the story’s core info 5. Burying is not being used enough ! Second Revision ! What’s changed? 1. 2. 3. 4. Gradients reduced, bevels gone Metadata reduced to a single line, no visual jump Share replaced ‘blog it’ ‘email it’ Bury has single-click option and prominence Success? • • • Buries increase Testers find chronology slightly clearer There’s less stuff Digg Comment System 2005 2006 2007 soon... First Design, 2005 First Design, 2005 Challenges become obvious • • • • Design is awfully literal Data isn’t very compact Limited Functionality Doesn’t scale Second Design, 2006 Second Design, 2006 But... still doesn’t scale • Nesting encourages discussion but now there are more comments than ever Second Design, 2006 Nesting confuses novices • • • User testing reveals confusion Don’t understand ‘indented comments’ Top-posting adds to the problem Second Design, 2006 Too little nesting for experts • • Users implicitly say this by using @soandso to fake functionality Users explicitly tell us by email and comments Second Design, 2006 Content organization... • Block button is too abstract and too cluttered Third design, 2007 What’s changed? 1. Header is simplified... less image heavy 2. Block button hidden 3. Threads collapsed a. Scales better b. Lessens top-posting c. Simplifies initial view Explicit feedback What are people saying? 1. I liked it the old way (you should be fired) 2. Annoying to open threads (too slow) 3. Hard to find your own comments 4. Ugly threading, too many vertical lines Implicit feedback What are people doing? • • • Total comments increased by over 30% Unique commenters increased by around 20% Unique comments per person increased by about 15% Successes & missteps ! Top-posting is lessened and doesn’t confuse novices ! Scales to thousands of comments ! More people are participating ! More comments are being posted ! People are threading better Successes & missteps x Too slow to load individual threads x Design is cluttered x Especially when threads are open x Can’t find your own comments if threaded x Interactions are poorly designed What to do about it... • • • • • • Go back to the old design? Address shortcomings Adapt to user feedback Get more feedback -- we released a video User test this time Leave time for further iteration What to do about it... • • • • • • Go back to the old design? Address shortcomings Adapt to user feedback Get more feedback -- we released a video User test this time Leave time for further iteration Let’s see it... Pownce Initial Design Pownce is brand new... • • • No established design language No prior expectations from users Blank slate is intimidating though Inspiration leads the way First take is close... Focus on the note... Focus on the note... Hatch adds movement & character Visual language spreads An ‘aha’ moment. Leads to more things. A few conclusions • • • • • Starting focused can be a big help Echo design language Be prepared to iterate and react to your users Many types of feedback Implicit feedback is especially important There’s nothing wrong with making mistakes Thanks! Slides will be on Slideshare pownce.com/dburka or deltatangobravo.com Illustrations by Ryan Putnam (via istockphoto.com)