June 30, 2000

7:34 AM

Has Amazon.com finally ditched the unsustainable and cluttered one-tab-per-store interface? And if so, is this how they react to their stock falling?
posted by tranquileye (12 comments total)

Huh? The site looks the same as ever to me.
posted by daveadams at 7:50 AM on June 30, 2000

This is just a test from amazon of a new interface. They have a page up describing it here. I'm not quite sure what to make of it yet. We've been analyzing it all week here at my office.
posted by jbelshaw at 7:51 AM on June 30, 2000

Am I the only one who thinks that Amazon's springing new redesigns on selected random visitors periodically is sort of a FUD tactic? Especially when they don't get used in the end, as in the portally test design from February? Maybe it's just me, but I feel like if I were in charge of a website of Amazon's size & traffic, I wouldn't be messing with the user experience at all unless I was committed to it. And I haven't hit the jackpot yet trying to get served up the newest attempt, but that last one, unless I missed it, didn't have anything at all to say "here's our new look, here's why it's different, what do you think, here's a link to the old look for those of you who want to shop as you are accustomed rather than being guinea pigs".

I'd be curious to know if any MetaFilterites have been involved in a redesign test conducted like this, what the rationale was, and how it went.
posted by Sapphireblue at 8:07 AM on June 30, 2000

Oh, how I love the hubris that's dripping from this statement. As if Amazon itself invented the navigation tab. As if navigation tabs were such a great idea in the first place.

As for their "new navigation", it may be an improvement. But it's still the fruit of a poisoned tree.
posted by ratbastard at 8:12 AM on June 30, 2000

I remember about 6 months ago, Netscape was randomly serving one of 3 redesigns to visitors to home.netscape.com and asking them for input. A random live test isn't a new thing.

(BTW, fixed that blockquote up there)
posted by Electric Elf at 9:11 AM on June 30, 2000

Thanks, elf!
posted by ratbastard at 9:16 AM on June 30, 2000

I didn't say it was new. I did wonder if it's a good idea :>
posted by Sapphireblue at 9:16 AM on June 30, 2000

Serving unique version of a site to selected visitors makes a lot of sense in some cases. I don't remember exactly how amazon did it, were they giving each user a different version at random or was it only a small percentage of hits? I know for a fact that deja.com-- chime in any time Lane-- uses a tool that allows them to serve a testing version of their site to a predeterminded precentage of visitors, say 2%. In the grand scheme of things even if those 2% have a horrible experience (which they wont, because none of the redesigns that amazon has tested were that horrendous) it will probably be outweighed by the data collected.

It's important to note that Amazon may be looking for more than straight usability 'stats.' I'm sure marketing ("did more users click into the DVD store with this design?") is a large part of it also.
posted by bryanboyer at 9:44 AM on June 30, 2000

yeah, the deja.com tool was wildly underused but pretty damn clever-- it would redirect a given percentage (say, 2%) of the deja.com populace to a different interface set, and an equal number to a "control" group of the standard interface. this was done via dividing the GUID for each user, set in the cookie, by a factor of something, resulting in a remainder between 1 and 100 -- so you'd redirect everyone with a remainder of, say, 1,2,3, and 4, and that was 4% of the day's traffic.

run the test for a week or so and then you could do all sorts of cross-comparisons between the two sets, macro and micro. an amazing, interative interface development tool. they probably should've just marketed that, they'd've done better. :)

the problem deja.com had was that it was a useful tool for interface design issues, esp. small ones (i.e., should that link go on the top or the bottom?), but sucked for larger things (like which overall site design was better.) people with agendas were always reading more into the data than was actually there. people who shouldn't have even been given access to the data. rather annoying.
posted by monstro at 9:56 AM on June 30, 2000

How about a MyAmazon type deal, where the user can choose (or be automatically given) their most frequently used tabs?
posted by EngineBeak at 3:34 PM on June 30, 2000

Ha ha, hadn't even seen this yet; "MySlate" -- a new setup to select articles for a single serving
posted by EngineBeak at 5:15 PM on June 30, 2000

A "My Amazon" feature would be nice, from a user's perspective and definitely from a usability standpoint. But if I was selling expensive leather-bound DVD cases, and a user wasn't able to see that case because they don't have the DVD store in their My Amazon, I'd be peeved.

For users, it'd be great. For businesses, it'd be awful.
posted by hijinx at 8:50 AM on July 1, 2000