Comments on: Primary & Secondary Actions in Web Forms

By: tommy australien

tommy australien — Thu, 01 Nov 2007 07:32:31 +0000

Neat article, but as for the examples I’d agree with em_te. Get rid of the “cancel” button altogether as in 99% of all cases it is not needed. Simply because everyone else does it, does not mean you have to do so also, something that struck me a while ago.
Consequently I redesigned a few froms and projects, reducing the number of buttons, which may have been redundant or missleading.
When it comes to usability, less often is more.

By: Richard Davies

Richard Davies — Thu, 06 Sep 2007 16:29:51 +0000

Where was this post last week when I was desperately searching the web for information like this to help me out with a current project? ;-)

By: Barnaby Claydon

Barnaby Claydon — Thu, 06 Sep 2007 03:13:26 +0000

The requirement for “Cancel” is entirely dependant on the type of application. We’re not really privy to the user flow in this study with only this one page being used. You would have to give the users an indirect reason to utilize the “Cancel” button to be able to /fully/ assess the usability, but one a simple sign-up page such as this 99% of all web users would simply click their browser’s back button; “Cancel” only comes into play in the midst of a multi-step flow.

Anyway, I thought the differences in alignment were telling although somewhat predictable.

By: Sensei Loco

Sensei Loco — Wed, 05 Sep 2007 16:29:27 +0000

My first reaction is that the buttons’ order should be swapped. Cancel -> Esc, Submit -> Return
Then again, I kinda agree that the Cancel button prob shouldn’t even be there.

By: Searle

Searle — Wed, 05 Sep 2007 11:19:16 +0000

I don’t think the test is thought out well. This is why:

- Users are used to recognise grayed UI elements as “not available”. Using gray is wrong. A lighter color would be better, maybe.

- Tests with images are missing completly. Placing e.g. a “Arrow to the Right” icon next to “Submit” (but inside the button) makes the button intuitively primary, in my opion.

By: Justin L

Justin L — Wed, 05 Sep 2007 09:30:43 +0000

I see it’s fashionable at the moment to use plain links for cancel, or secondry actions – ala facebook, basecamp. but why are these secondary anyway? The actions themselves are pretty black and white

By: Will Peavy

Will Peavy — Wed, 05 Sep 2007 02:29:52 +0000

From a usability standpoint – I think it is best to not use buttons that say Cancel / Clear / Reset … etc. These types of buttons are rarely intentionally used, and are more likely to be accidentaly clicked in place of the Submit button.

By: Ian

Ian — Wed, 05 Sep 2007 01:31:54 +0000

@dwhsix: Not a contradiction really. Option B was the quickest from start to finish. This is because users just presumed the placement of the buttons at the bottom, automatically clicking on the left, ignoring the right. This is backed up by the failure in option E where the buttons are reversed.

As option A and C had a different visual element (cancel), the user had to slow down to understand the function of this rogue visual element. As such, these designs help ensure that the user clicks the correct button.

So, B is quicker but A and C are more accurate. This is what Luke is getting at.

By: bmiedlar

bmiedlar — Tue, 04 Sep 2007 20:09:01 +0000

It would be interesting to see if right alignment would have equal success if the buttons were reversed in order.

By: dwhsix

dwhsix — Tue, 04 Sep 2007 18:53:56 +0000

Luke’s article and study looks very impressive, with lots of fancy graphics and ‘fixations’ and heat maps. But if you read it closely, the empirical evidence contradicts the claim he puts forth.

He claims

…the best practice Iâ€™ve advocated has been to visually distinguish primary and secondary actions … Reducing the visual prominence of secondary actions minimizes the risk for potential errors and further directs people toward a successful outcome.

But from the data they collected, they concluded:

Option B performed best of all. Fixations were shorter and fewer in number when using this design. And people were also able to complete the task more quickly and efficiently than they did when using the others.

What was Option B? The one where the secondary action had, in fact, just as much weight as the primary action (same button style, left justified).

By: Jim

Jim — Tue, 04 Sep 2007 18:00:57 +0000

Yes – it would have been interesting to just remove the “Cancel” button/link altogether and see how users reacted.

By: em_te

em_te — Tue, 04 Sep 2007 16:03:47 +0000

What if someone actually wanted to “cancel” in design A. Would they be more likely to click the link or hit their browser’s back button? Maybe that’s why design A is so good. Because the user begins to see in buttons and completely ignores links, making the link useless, which means the user really only has to deal with 1 action.

By: Liming

Liming — Tue, 04 Sep 2007 14:36:20 +0000

Very interesting topic. When we designed our software, our designer went through similar case studies, wish we had this discussions back then.