Hacker News new | comments | show | ask | jobs | submit login
Your Users Won't Read (w2lessons.com)
101 points by mwbiz on Jan 29, 2012 | hide | past | web | favorite | 23 comments



I've building my own site for the last 5 months now and I've been knee deep in usability like this. It always comes down to what is the easiest path you can possibly give to your users.

I'd do away with the whole extra click of making someone 'accept' or 'decline' the invite. Just have the two buttons in the dialog, on the first screen. Don't make me read, then click, then read again, then click. There is no logical reason for that extra step at all.

The button on the right should be the 'continue' or 'ok' button and the one on the left should be the 'cancel' button. Don't use text for 'cancel' or 'decline', just make them buttons. The buttons should be right aligned at the bottom of the text. This is how the user expects things to be based on UX experience of their operating system, don't confuse users by changing it up.

I just spent 5 minutes in Balsamiq mocking something up for you [1]. I'm not super happy with the decline button, I might play around with it in different areas (or ideally get rid of it all together). The point being that there is no need for the 'accept' step, since they aren't accepting anything, they are signing in or creating an account.

By the way, for my site, we got ride of the whole create an account process entirely. We just have sign in. User clicks sign in and they have the option of using their FB account (with the bare minimal permissions) or using a BrowserID account (which only requires an email and password).

Keep it as simple as possible.

[1] https://img.skitch.com/20120130-dh9hf6mtt8wwemtgcjy7rq16t2.j...

Good luck.


I work at a movie theater. We have four pairs of front doors, all of which open outwards. One pair has handles; these are the doors which should be used to enter the building. We sell tickets outside, and tear tickets at the door; if both these doors are open, people come in three or four at a time and we can't get to them and take their tickets. So, one of these two doors is always locked, and the entrance door which is not locked has a sign on it that says "Please use this door."

A nontrivial number of people will come up, read the sign, and promptly try to open the OTHER door, see that it's locked, and then walk away to try and find the "other door" - generally trying to push open the out doors. As it turns out, they all read the sign as saying "Please use other door".

(Occasionally they'll try to claim that the sign is wrong. There's a certain schadenfreude in the looks on their faces when I convince them to reread it more carefully.)

So, we removed the handle for the door that's always locked, and took down the "Please use this door" sign. Now people just walk along the out doors, pushing on them to try and find the one that'll open.


Putting a 'Please use this door' on a door that is open is redundant. That is why people 'believed' that the sign said 'Please use other door.'

If you were in front of the door that is open, why say 'Please use this door.' They will. You don't have to say so. People generally do not walk up to the door they do not tend to use. Hence why they assumed that the sign said 'Please use other door', since that is 1) what is most often on the doors and 2) it's illogical to have a confirmatory sign on the open door.


How about an 'entrance' sign on the door, and a red circle with a slash (no entry) on the other doors? Negatives tend to be less effective than positive statements, especially if the signs around are similar.


it's not so much that users won't read, it's that they read as little as they think they can get away with to make a decision. the best way to solve this problem is to take the decisions away from the user. instead of asking if they want to sign in, give them a username/password form. most of your users will probably be able to fill out the form. no decisions, no reading necessary, no problems.

the users that aren't able to fill out the form they will be unable to continue until they read. they will start looking for the "sign up" button. make it big and obvious for them. still, nobody has made any decisions. the user takes the only path available to them.


experience in support:

'The software isn't performing the analysis'

"Okay, what does the dialogue box say"

'The data file is corrupted' (we have no such dialog box)

"Hrm, I haven't seen that box before, can you read it out word-for-word to me?"

'The. Data. File. Is. Corrupted'.

"Hrm, I've not heard that before, could you please spell out the words letter-by-letter so I can get it exactly right for the developers to look at?"

'Sure: Y-o-u-r s-e-t-t-i-n-g-s a-r-e n-o-t c-o-n-f...'

Even when specifically asked to read the message out word-for-word, endusers find a way to screw it up.


Send that to NotAlwaysRight, now.


4 or 5 years ago I made a registration workflow for a event I was working with as part of a freelance gig. We had a giant bolded text link, something like 24 points, that said "Click Here to Register," right on the front page above the fold.

I got several emails from users who asked me "How do I register? I looked all over the front page and found nothing."

In hindsight, I wonder if maybe I needed to make it look like a button, or if these particular people were beyond saving and wouldn't have bothered to read the text on the button anyway.


The industry has trained people for a decade that primary calls to action look like buttons. People scan for this as a learned behavior, because scanning accomplishes their goals faster (true for most HNers), they don't enjoy reading, they can't read well, or they've learned by painful experience that words on a computer screen are likely scary gibberish like "screen resolution" and "defragment" and you should just click the blue thing in the bottom left corner followed by the blue E so you can get to your Googles.


On top of this, users have also been trained to ignore anything that looks like a banner or ad. Sometimes giant text just looks like something meant to be ignored.


You should see this website:

http://airforcefcu.com/


I didn't even need to expend any mental effort to ignore the large advertisement-shaped blocks there.

AdBlock Plus did it for me, and delivered me a blank page.


Apparently with no scripting, the middle of the page does not load.


It happened to me once when reading a FAQ for rootnode support IRC channel. There was a most important "point 0" in FAQ, that was emphasized by moving it above the page header. It was supposed to be the very first thing visitor will read, but my mind didn't even register this text being there (even after being told that there is a "point 0" and revisiting the page). Needless to say, I got into some trouble because of that, but I managed to convince FAQ maintainers that banner blindness[1] is a real thing, and they fixed the FAQ.

http://en.wikipedia.org/wiki/Banner_blindness


This article comes to the wrong conclusion. Removing the dialog wasn't what stopped the confusion — changing the text on the buttons did.

Your users will read — but only enough to feel they have an understanding of a situation before they take action.


I believe that the key word here is feel they understand. That does not necessarily mean that they actually understand.


Why is this a surprise. Reading is sequential. The primary way is which we consume information is not. Our eyes capture a bit of everything, left and right, up and down, then our brain gives it a meaning. Reading requires us to follow a specific path, and I am forced to believe it is harder on the brain.


A related study into how long people spend reading words on a web page found that on average people only read 20-28% of the text. For 1000 words, the average visit duration was less than 70 seconds.

http://www.useit.com/alertbox/percent-text-read.html

I read quite a lot of text online, and I certainly have a different style of reading a web page versus reading a book. I'd be fascinated to know how efficient scanning is after years of practice online, in terms of information consumed that you're able to recall. 28% is surely too low for adequate comprehension, though.


This is Apple's HIG stuff, from the 1990's.


In my experience. It's Murphy's Law adapted to reading.

If you pour a lot of care and attention in to some copy or instructions. They won't even not read it, they will flagrantly disregard it.

If you fluff up some filler text to fit a space, they will go through it with a fine tooth comb and hold you to the fire over it.

In my experience, you can't win.


If I ever get popup boxes on sites (or Windows), it becomes instinct for me to click the X or the button that says 'No', because its in the way of what I was doing and I don't care about whatever it has to say (if I wasn't expecting it).


"Don't make me think" is a pretty good book on this topic.


This is really good information. We did do some level of usability testing but we still rely on content for important messaging.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: