546524 - Implement Design A for IE.html A/B Test

Status: VERIFIED FIXED

(www.mozilla.org :: General, defect)

Description

[Laura Mesa [:lmesa] [:lvmesa]]

Please Implement Design A for IE.html A/B Test.

You can find the design here:http://www.sendspace.com/file/d7bjf9

The idea behind this page is to put all content above the fold and to limit links off this page to menu bars, the download button and navigation.

Comment 1

[Steven Garrity [:sgarrity]]

We'll get to work on these. I wondered if you could share a bit of the methodology behind these variant designs. It strikes me that the variants are different in so many ways that it would be difficult to garner meaning from the results.

For example, if one design A has the download button in one place, and design B has it in another, but the total file size of A is 3 times that of B due to all of the content/features - how can we distinguish the effect of download button position vs. loading time?

Also, what is the process for creating these variations?

Comment 2

[Laura Mesa [:lmesa] [:lvmesa]]

(In reply to comment #1)
> We'll get to work on these. I wondered if you could share a bit of the
> methodology behind these variant designs. It strikes me that the variants are
> different in so many ways that it would be difficult to garner meaning from the
> results.
> 
> For example, if one design A has the download button in one place, and design B
> has it in another, but the total file size of A is 3 times that of B due to all
> of the content/features - how can we distinguish the effect of download button
> position vs. loading time?

This is a great point. Load time may definitely be a factor...as can many things. I think it will be hard, because of so many differences with the control, to be able to call out exactly why one design performs better than another. It may be a combination of factors.  The idea is to see which design prevails as a whole. Once we test that, then we'll make tweaks to the "performing' design to try to optimize performance as best we can. It's probably during this process that we'll start to tease apart the design factors that made the 'performing' design 'win' in the first place. 
> 
> Also, what is the process for creating these variations?
Right now we go through pages we like (southwest airlines, google, itunes, etc) and see how we can try and adapt those designs to our needs.  It's also very subjective, adding and moving things based on taste, etc. 

Does that help? I'm happy to talk more over the phone if you like.

Comment 3

[Steven Garrity [:sgarrity]]

This page is set up in trunk in r63171.

Preview here: http://www-trunk.stage.mozilla.com/en-US/firefox/ie-a

Comment 4

[Laura Mesa [:lmesa] [:lvmesa]]

Looks awesome, Steven!  Thanks!

Comment 5

[Blake Cutler]

Can we push this to production?

Comment 6

[Alex Buchanan [:abuchanan]]

r63222 on stage
r63223 on production

Comment 7

[John Slater]

Page is live and looking good - thanks all!

Before we resolve this, I have one small tweak: Steven, can you increase the white space between the main content and the footer a bit? Right now it feels like the footer is crowding things...since we want to test what a much simpler page would be like, I'd rather not have that footer so close.

Thanks much!

Comment 8

[Steven Garrity [:sgarrity]]

(In reply to comment #7)
> can you increase the white space between the main content and the footer a bit?

Sure - I had wondered about this myself. Done in trunk in r63270. Merged to Stage in r63271. One file to push to production: style/tignish/ie-a.css

Comment 9

[Laura Mesa [:lmesa] [:lvmesa]]

Alex, John?  I've been trying to push this and every time I do I get an error message. Thanks!

Comment 10

[Alex Buchanan [:abuchanan]]

pushed.

Comment 11

[raymond [:retornam] (inactive)]

verified fixed on http://www.mozilla.com/en-US/firefox/ie-a
and http://www.mozilla.com/en-US/firefox/ie-b
 
in IE6, 7 and 8