7. Card Sorting is as common as Lab based
Usability Testing
Source: 2011 UxPA Salary Survey
8. Terms & Concepts
• Open Sort: Users sort items into groups and give the
groups a name.
Closed Sort: Users sort items into previously defined
category names.
• Reverse Card Sort (Tree Test) : Users are asked to
locate items in a hierarchy (no design)
• Most Users Start Browsing vs Searching: Across 9
websites and 25 tasks we found on average 86% start
browsing
http://www.measuringusability.com/blog/card-sorting.php
http://www.measuringusability.com/blog/search-browse.php
11. Method & FAQs
• We want to improve the Target.com navigation.
– Obtain a baseline set of metrics and problem spots
– Make improvements
– Compare against baseline
• Pick a variety of items to test the taxonomy
• Conduct an Open Sort then Tree-Test
• Use Product Names with some level of context
12. 40 Items Across Departments
1. Presto GranPappy Electric Deep Fryer
2. Black CleanAirBall Air Purifier
3. Nano Tech Shaver
4. Hamilton Beach FashionFirst Medium Iron
5. Brother International Sewing Machine-
XR9500PRW
6. T3 White Micro Featherweight Hair Dryer
7. Signature Sleep Silhouette 10” Queen
8. Thomas O’Brien Canterbury Blue Bath Rug
9. Fieldcrest Luxury Towel Collection
10. Privet House Sugar Sack Pillow
11. P Dog ZIP Lead Retractable Blue Leash
12. Kitchen Aid 3pc BBQ Set
13. Women’s Mossimo Kasey Boot
14. Women’s Mossimo Pearce Black Leather
Pumps
15. Men’s Brown Swiss Gear Jackson Hiker
16. Totes Clear Bubble Umbrella
17. Merona Yellow Scarf Tote
18. Wildkin Kaleidoscope Backpack
19. NBA Golden State Warriors Blue Water Bottle
20. Grey Stripe Merona Men’s Short-Sleeve Polo
21. Corellian Freighter Men’s Graphic Tee
22. Inset Mother of Pearl Stainless Steel Cuff Links
23. Xhilaration Black Large Sunglasses
24. Wrangler Mens Reversible Belt
25. Cherokee Boys Drawstring Cargo Short
26. Girls Danshuz Pink Short-Sleeve Sparkle Straight
Dress
27. Tiddliwinks Blue Dot Plush Monkey
28. Black Griffin Case for Apple iPod touch
29. Marvel POWER ATTACK AVENGER: IRONMAN
30. Titanic (2 Discs) (Blu-ray/DVD)
31. PC Diablo 3 Computer Game
32. PC Treasures FlipIt 10” Tablet Sleeve
33. Commodore Big And Tall Executive Oversize Chair
34. Classic Borders Wedding Invitation and Response
Card
35. Hello Kitty Party Favors 32-pc.
36. Women’s Obsession by Calvin Klein Eau de
Toilette
37. C9 by Champion Women’s Molded Cup Racer
Back Sports Bra
38. Coffee Shop Outerwear Women’s Plus-Size Spring
Twill Coat
39. Women’s One Shoulder Shantung Dress
40. $5-$1,000 Iconic Puppy GiftCard
13. Overview
• 50 Participants
– 65% Female
– 2 Excluded for Quality Control (Haphazard Sorts)
• Sort 40 Items into 10 Fixed Categories
• Median Study Time: 11 Minutes
• Output
– Dendrogram of Clusters
– Open Category Names
– Most Difficult to Sort
14. 12 Clusters
1. Men’s / Boys
2. Women’s / Girls
3. Accessories
4. Runt: NBA
5. Kitchenware
6. Bed + Bath
7. Runt : Furniture
8. Toys
9. Electronics
10. Runt: Shaver
11. Misc/Pets: Gift Card & Leash
12. Misc: Party Favors & Invites
16. Most Popular Category Labels
Similar Category Names Combined
Percentages are % of 48 participants who labeled a category, so % adds up to more than 100%
Error Bars are 90% Confidence Intervals
17. Most Difficult Items to Categorize
Percentages are % of 48 participants who selected item, so % add up to more than 100. Error Bars are 90% Confidence Intervals
22. Selected 14 Items From Open Sort
1. Wrangler Mens Reversible Belt
2. Marvel POWER ATTACK AVENGER: IRONMAN
3. Women’s One Shoulder Shantung Dress
4. Grey Stripe Merona Men’s Short-Sleeve Polo
5. Inset Mother of Pearl Stainless Steel Cuff Links
6. Women’s Obsession by Calvin Klein Eau de Toilette
7. Tiddliwinks Blue Dot Plush Monkey
8. Hello Kitty Party Favors 32-pc.
9. Nano Tech Shaver
10. Brother International Sewing Machine-XR9500PRW
11. Classic Borders Wedding Invitation and Response Card
12. Kitchen Aid 3pc BBQ Set
13. Men’s Brown Swiss Gear Jackson Hiker
14. Wildkin Kaleidoscope Backpack
23. Overview
• 50 Participants
• 68% Female
• Median Study Time 17 Minutes
• Task Based Metrics
• Confidence
• Difficulty (SEQ)
• Success
• Time
28. Brother Sewing Machine 64% -The sewing machine didn’t have a separate section that I could find
- it could fit into a lot of categories and it made it hard to nail down
Wildkin Kalidescope Backpack 38% - Not sure if it’s a child’s toy or real camping gear
- Expected to find it in the children’s section
Hello Kitty Party Favors 36% - I didn’t find a section for things like Cards or Party Favors Could be for
both men and women
- not apparent under what main heading it would be under.
Men’s Brown Swiss Gear Jackson Hiker 34% - Did not know what this was, a boot? A Jacket?
Classis Border Wedding Invitations 39% - Was expecting to find a stationary or cards section
Twiddledinks Blue Dot Plush Monkey 12% - Did not know if this was a toy
- No Stuffed animal section
Kitchenaid 3-pc BBQ set 10% -Did not know if it was Kitchen or a small appliance
- Could be categorized under kitchen or outdoors
Marvel Attack Avengers: Ironman 8% - Did not know what kind of Toy it was
- List of Toys was too long, it was hard to eliminate
Nano Tech Shaver 6% - Is it for a man or a women?
- Did not know if it was electronics or another category
Women’s Obsession by Calvin Klein 6% - Did not know which category to put it under
Mother of Pearl Stainless Steele Cuff links 2% - Could have fit into more than one category
Item % that found it difficult Reason it was difficult
Verbatim Reasons For Difficulty
29. How difficult or easy was it to complete the task
Very EasyVery Difficult
Percentages are normalized percentile ranks from across 200 tasks. 50% = Average.
32. Hardest Items to Categorize – Tree Test Vs Open Sort
Difficulty sorting explains
about 16% of difficulty
finding.
33. Sample Size Considerations
Margin of Error
90% Confidence (+/-) Sample Size
24% 10
20% 15
15% 28
14% 32
13% 38
12% 45
11% 54
10% 65
9% 81
8% 103
7% 136
6% 186
5% 268
4% 421
3% 749
See Chapter 6 in Quantifying the User Experience
34. Conclusions
• Determine if new categories should be created
or consolidated
• Look to confirm or deny hypotheses about
categories and findability
– Gender Searching
– Sub Categories
• Watch for “false” sorts
• If “key” items are hard to find, move them,
test them
• It’s hard to do a Card Sort without a Tree Test
On this bridge in Sarajevo in June 1914, the Archduke Franz Ferdinand of Austria was assassinated. His death set into motion the first World War.
A few weeks earlier, the May issues of Science published an article by Robert Yerkes on studying human behavior in the burgeoning field of psychology. Much of Yerkes work on human intelligence would be applied as millions of US men were drafted and sent off to fight the War to end all wars. In that same issue of Science was another article on the use of Cards as the ideal tool for psychological measurement.
Card sorting as a method in HCI largely took off during the internet boom of the late 1990’s with the proliferation of website navigation.
Today it’s one of the most popular methods UX professionals use. In fact, practitioners report using Card Sorting as frequently as task oriented lab-based usability testing.