How to categorize usability/ux problems?

Question

Even though the definition of usability/ux problem can be elusive I would like to learn more about how to categorise usability problems. And by categories I mean various qualities of the user interaction, not a quantitative categories like "severity" or whether problem has workaround or is persistent.

So far I found few fragmented information like Usability Problem Categories based on Jakob Nielsen's Usability heuristics (and original) or How do usability specialists categorize usability problems?

Is there any industry standard for usability/ux problem categorisation?

Answer 1

The Kano Model allows us to fit user needs into three categories; from this you can match your usability problems to what type of feature they impact and prioritize from there. From Wikipedia's article on the Kano Model:

Basic needs are things people need to do, this is where poor execution is really bad and amazing execution doesn't sound out as much. You notice these when it's broken, so they're a good priority for usability issues. Think about email; most people don't want an amazing email client, they want something that sends messages reliably.

Performance is where efficiency really counts: when they're not met, it's really bad. When they're met exceedingly well, they're great. These need a large amount of attention. Think Microsoft Word; every bad thing a user experiences in Word affects their performance. Even having a menu in the wrong place can ruin a worker's productivity.

Excitement is something that doesn't have a negative impact when it's not performed (or isn't performed particularly well) but it's inclusion has an immediate, joyous impact on it's users. Think Flipboard. There is no practical purpose for the flipping animations, but it creates instant amusement which is tied directly to the brand.

The Kano Model categorizes features not problems with features, but by knowing which features and what sort of problem (ugly design, poor efficiency, ) you can better assign priorities and severity to specific issues.

It's not quite an "Industry Standard", there is no Diagnostic and Statistic Manual (DSM) for UX/Usability, but it's a proven effective way to categorize user needs.

Answer 2

What are you attempting to measure? User satisfaction? Web site effectiveness? Or simply a meta-study about usability issues? Only the third seems to have any merit (in my limited understanding) for any sort of "non-quantitative" analysis, so I'm anxious to see what folk posit for your post :)

Also, what is the product of the study? I guess I can presume a web site for the purpose of generating discussion and helping you find some starting ideas.

You could start by identifying common usability errors and generalizing them:

feature disclosure

• user can't find feature needed to complete task
• user could not navigate submenus
• user expected feature to be at x but it was at y

lack of affordance/indication

• user can't tell link is clickable
• can't tell button is enabled
• didn't know element was draggable
• didn't find table's right-click context menu

obscure mental model

• user makes incorrect assumptions about the program's behavior or architecture
• the functional model doesn't match user's assumptions or goals
• the help instructions misled user about how program functions

blocking mode/state

• user could not access feature x in mode y
• user could not figure out why link/button/feature was disabled
• user thought it was okay to do z, but program prevented this (could also be obscure mental model)

I'm sure it wouldn't be hard to come up with plenty more by taking this approach.

EDIT

Upon review, this seems almost identical to the approach Jacob Nielsen utilized, so it must not be all that bad.

Answer 3

I prefer to categorize problems based on the area of UX practice that is necessary to fix them (e.g. information architecture, ergonomics, visual design, process design, copywriting, etc.) plus related disciplines of engineering/manufacturing (build quality) and business needs (if any unethical manipulation is spotted). This way I can quickly direct tasks to the person who can perform them the best.

Under such an approach, Nielsen's "visibility of system status" can be any combination of:
- process design (at which stage does the user need to know of the system status?);
- information architecture (what kind of a system status is needed to be shown to this user?);
- ergonomics (where should the system status be located?);
- visual design (how should the system status look?).

Or "error prevention" can be any combination of:
- process design (is there any fail safe logic in the process?);
- copywriting (are the directions clear and do they set the right expectations?);
- ergonomics (is the product designed to fit mental models and behavior?).

Then everything is scored on subjective level of problem severity to prioritize the order, in which they will be addressed.