Usability Recommendations

[#1]

Usability engineers & researchers are crucial parts of User Experience. While not as "glamorous" as designers they bring evidence to the world of design. Through observation of heuristic tests they offer empirical evidence that a design is working, failing, what users like, what users aren't finding, what users are doing that they don't even realize they are doing it, etc. Usability is the closest thing design has to being a science.

The attached technical paper goes through a variety of topics with usability in mind and makes recommendations. It links out to studies and research already done supporting best practices.

The basic findings and recommendations in this document are that:

• People don’t read everything online, they skim

• Paragraph width impacts comprehension

• Scrolling is good

• Monitor Resolution: design for 1024x768

• Ipads: design for 1024x768 and allow scrolling

• Icons don’t necessarily help usability

• Filters should be on the left

[Or]

Hello Michael,

I have to say I was happy to see this post - usability guidelines are something that QlikView lacks. Unfortunately, having actually read through he paper, I came out a little confused. Virtually everything in the paper seems to apply not to analytical platforms, but to web pages. QlikView applications are most decidedly not web pages, and what may apply for one does not necessarily apply for the other.

To point out the most surprising example, "Scrolling is good". In explaining why scrolling is good, only one citation is made - to a 15-year-old paper that refers specifically to web pages. Yes, scrolling might sometimes be better than fragmenting related information into several sheets, but that doesn't make it good - just the lesser of two evils. Quoting from Stephen Few's "Now You See It" - "We should avoid fragmenting information that we're examining by placing it on separate screens or in locations that we can't see without scrolling" (page 51, emphasis mine).

I have found no body of research suggesting that scrolling is superior to using tabs (or, in QlikView's case, container objects), and my own experience suggests that users are more likely to notice container objects than they are a scrollbar. In both cases the ability to make comparisons using our working memory is lost, but at least with the container objects, the header of each available chart is immediately visible, meaning the user is aware of it. Anything placed under the fold may as well not exist, unless the user knows to look for it there (if you regularly have scrollbars, a user would check - but if that's the case your overall design is probably not great).

I may be entirely off-base here - in which case I'd love to see some references to the contrary - but the document does not list anything that appears pertinent to the question of designing an analytical tool rather than a web page full of text.

Similar issues exist with several other points such as  "People don't read online" and "Paragraph width impacts comprehension". In both cases, these are true - for web pages. We're not dealing with web pages, we're dealing with a Business Discovery platform. If you're displaying paragraphs of text, you're almost certainly doing something wrong to begin with. The best policy for paragraphs in QV is "Don't".

I am happy to say that, unlike the above points, the back half of the paper is sound and QV-relevant. I would add, for tablet / smartphone design, that listboxes need to be significantly bigger on a touch device than they do on a computer - our fingers are much stubbier than our mouse pointer. Maintaining your computer-monitor design for a tablet would force either an unreasonable amount of zoom/unzoom for making selections, or force users to use a stylus - neither option would be popular.

So, overall - I'm glad to see QV taking additional steps into the realm of usability and design - but I'm having a hard time swallowing some of the recommendations in this document without relevant sources and citations.

[rwunderlich]

Hi Micheal,

Nice to see a post on this topic. Brings up some interesting discussion points.

Although we can apply some web page principles to QV design, QV is different from webpages in some very important ways.

- Users are almost always "scanning", very rarely "reading".

- QV does not support floating objects.

The last point is why I strongly disagree with the "scrolling is good" recommendation. When you scroll a QV sheet, important contextual information such as Current Selections box and Listboxes scroll out of view. Even when not scrolled completely out of view, objects become partially visible creating a ragged and jagged viewing experience.

-Rob

[paulyeo11]

nice article. i agree the first . qv does not provide usability guide line.

last month i bought a lenovo yoga. it have touch . and agree scrolling is good, other then that i still have issue on zoom in and out. which is i still dont have solution. may be can some one share with me.

[Anonymous]

Rob,

Great point about floating objects.  That's always bothered me.  That seems like something that should be so simple to implement.  I imagine that it's doable with custom objects in AJAX, but it would be much better if QV natively supported it.