How do you decide when a website or application is “usable enough”?

Question

I did see the related question

How do you know when an interface is complete - Though the answers in this question are very interesting,they are more from a product management perspective of what to keep and what to ignore with regards to a finalized product.

But can we really determine when a product is finally usable ? Is there a defined parameter like "90 % of the users found this usable" or "All the associated heuristics which are applicable to this application are satisfied" or "All the potential use case scenarios were accounted for with the user using the optimal path as opposed to using the alternative route"

Answer 1

This is in fact your decision.

You must define the usability goals for the project in the usability specification. They are usually formulated just like you suggest: "n% of the users accomplished the task in less that n minutes". A couple of "standard measurement" have evolved during the last decades. Measures like time-on-task, success-rate, error-rate, efficiency and preference are pretty common. Take a look at "Measuring the User Experience" by Tullis and Albert. They cover the most essential issues in a very comprehensive and straightforward way.

The usability goals are set by looking at existing systems, competing systems or prototypes. In some circumstances you can set a goal that "sounds reasonable" (i.e. by following your gut feeling), but you should usually have some experience with similar systems and some basic performance- and preference measures as a basis for your new usability goals. Even if the goal is pretty obvious, it should still be explicitly stated in the usability specification.

Answer 2

1. Designing for a problem can have different solutions, say there is a proposal to create an "interface for an insurance company" - so the approach to find an answer is not one. On top of it - when if a interface needs to be deemed usable - then in a practical world when the design solution is made, its a judgemental call, where many do not map it to the final metrics or score-card to evaluate that "OK, here is an interface that is 90% usable, there is another one 95%" and pitch their proposal. Generally evaluators will look at the way "this has this problem than the other one"

2. But if it were "products - for an anti-virus software", the usability of the interface will undergo rigorous testing and internal evaluation with bench-marks. Purely it can be mapped through competitive landscape of similar products, emotive factors of the design, pure conventional mapping, pyschological or behavioral factors, information design validation. Once again, I feel you cannot address all that is available in the framework of metrics chart - but its important to make it tailored-made to particular problem or addressing towards the Key metrics of USP.

3. In non-ideal world, the Metrics or Heuristics can be relatively framed into various segments like Interaction, Information & Visual design etc., and Optimal solutions are bench-marked, and by a decision can be thought about on Parallel prototypes. Sometimes I do this way. But more often, when you do have multiple prototypes and want to find one usable - it becomes fairly a judgemental call. On practice and expertise, you know which works better, since your mind works to gather all metrics and quickly scans through which could be the appropriate option.

4. On the factual side, ground rules of testing a product with the users and stake-holders will drive to know the usability success. Surveys, polls and inspection methods help to bring in lot of foot-hold to the interface. Its also handy to have these tested reports for future modulation.

5. There is also a way where the metrics can be confined to "Time-based analysis" of task flow - where the usability is more skewed on time-rate.