[Sigia-l] Jared Spool's article on galleries

[Olly Wright]

This post is referring to Jared Spool's article on designing product  
galleries:

http://www.uie.com/articles/galleries/

As the lead IA responsible for designing the Sony Ericsson product  
catalogue, which Jared critiques in the article,  I thought I'd share  
some of the background context behind the design of the SE handset  
gallery and why it is the way it is.

I'll also say that we're very open to suggestions for improvements :)

I interpret the main point of the article as "You need to provide  
clear criteria for users to make their product selection".  The  
subtext is that product features is an appropriate way to do this.

In the case of the Sony Ericsson gallery users are given two prompts:  
the model name and a thumbnail size picture:

http://www.sonyericsson.com/spg.jsp? 
cc=us&lc=en&ver=4000&template=pp1&zone=pp&lm=pp

We have a problem, Jared states it clearly:

"However, they left out a critical component: the information that  
those users will need to make their choice. Unless the user is  
already intimately familiar with the phone they want, supplying model  
names and pictures isn't going to help."

I agree. It is a big problem. Here's some of the business context:

- Model numbers are arcane across the handset vendor industry.  
Unfortunately as IAs we have no ability to influence this. This is  
primarily caused by the fact that most companies produce a high  
volume of handsets with very similar, overlapping specifications.  
This prevents them condensing down into some limited sub-brandable  
product lines (eg. iMac, iBook, Powermac, Powerbook, Mac mini). The  
hardware just doesn't categorize neatly enough, nor is this likely to  
change in the future.

No handset manufacturers categorize their products. This is for a few  
reasons:

- Over the short market lifespan of a handset, they will move through  
several segments / 'levels'. A high-end (eg, gadget - geek - business  
oriented) handset at launch will be a mid-range model by the end of  
the year and a low end model the year after.

- Different handsets are positioned differently in different markets.  
A mid-range European phone is a high-end Latin American phone. The  
shift in 'level' I mention also happens at different points in  
different markets. The business / content processes needed to support  
this kind of categorization are highly complex. There are also many  
negative connotations to positioning a phone as 'cheap' or 'low-end',  
or 'limited-features'.

- Categorizing phones by lifestyle / consumer type (eg Fun, Serious,  
Simple etc) is fraught with problems. Not least, that phones just  
don't fit cleanly with lifestyles. There are businessmen out there  
who want very low-tech simple phones, and early-adopter consumers  
that want the latest wi-fi 3G smartphone the day it comes out. Also  
lifestyles have (in my experience) negative connotations with many  
users, who do not like to feel themselves being categorized, and  
generally perceive themselves to be a blend of multiple lifestyles.

To summarize: categorizing phones is right-out.

The next possibility, and one we have investigated in depth, is to  
provide short summaries of the main features, which users can use to  
make their initial selection criteria. Whilst this is intuitively a  
very promising route, it has a few major issues:

- The most interesting is that our research shows that the great  
majority of consumers do not make their handset selection based on  
features! In fact, emphasizing features is a turn off for them. This  
is a problem of the industries' own doing up to a point: by cramming  
a very large amount of technology into a handset they have managed to  
create the perception with the majority of consumers that phones are  
over-complex and contain unnecessary features. If you mention a  
feature in that context, or lead the description of the handset with  
features, you turn people off, not on. This does not apply to certain  
user groups of course, but since we can't categorize by user group...
It's similar to trying to explain the features of a PC to an average  
(non-technical) consumer by saying: 512MB RAM, 100GIG HD, 128MB  
Graphics card, S-ATA, AGP 8X, DVD-R. It's just not what they want to  
hear. This is why for example Dell describe a PC as 'For  
entertainment and productivity' (the merits of that approach I won't  
comment on here). iPod is not sold on features for example (whilst  
almost all other MP3 players are), other than the very simple  
messages contained within '1000 songs in your pocket' (ie 2 features:  
large storage capacity and small size).

- So what is the primary driver for phone purchase? It is how the  
phone looks. Essentially, the great majority of users will make an  
initial selection based on appearance and then delve further into the  
phone, trying to get a feel for its overall complexity / cost rather  
than zeroing in on individual features. This applies as much to  
offline (typically in a store) as online.

- Almost all phones have all the key features a consumer is  
interested in. When you user test, certain features do come up as  
important to consumers. However, nearly every phone has these  
features. Examples include: colour screen, integrated camera, music  
player, and for many users, whether the phone has games. The fact  
that nearly every phone has all these key features makes selection  
criteria based on these features meaningless. You can of course go  
more in depth (eg 2 megapixel versues 1,2 megapixel camera), but then  
you turn users off by being too technical.

- Again the global nature of the handset business causes issues. An  
important feature in one region is unimportant in another. US users  
are very interested in email for instance, whilst Europeans are much  
more interested in text messaging. Whilst this is theoretically  
manageable, managing key features on a per-location basis does add  
greatly to the costs of content management and ongoing maintenance  
overall. We have to weigh very carefully the financial and efficiency  
costs of the IA decisions we make with such a large site.

- On a more basic level, the addition of a significant amount of  
feature content / functionality within the product catalogue has some  
inevitable usability negatives (most in terms of a more cluttered page).

The user behaviour we observed with our product gallery is quite  
close to what Jared describes as Pogosticking. They will go to the  
gallery, click on a phone based on its appearance, and then proceed  
go back to the gallery and down into another product until they have  
looked at a selection of handsets. However, when asked, users have  
generally responded that they are satisfied with this process, and  
usually end up with a phone that they feel confident about.

I think partly this relates to what Jared states:

"We've noticed that users expect the most important items to always  
be listed first in the gallery".

This is the case with the SE gallery. We have some behind the scenes  
ranking going on that pushes the 'main' phones to the top. 'Main'  
means a lot of things, including different things to the local  
marketing people in different locations.

We also highlight the three main phones of the moment, and include a  
very short text snippet that attempts to sum up the phone. The  
difficulty of getting this copy right is huge however, and is no  
panacea. The big difference is that the pictures are bigger... simple  
yet effective in drawing attention.


I'm very open to any feedback on this, and particularly welcome any  
good suggestions for improvements :)

Olly Wright