Link List Color

Jakob Nielsen's Alertbox for May 13 is now online at:
  > http://www.useit.com/alertbox/link-list-color.html

Summary:
Lists of links are an intermediate case between content-embedded links and
menu items. Showing listed links in blue or in the site's main link color
is the recommended design - and the one most intranets follow.
	
----------------------------------

Usability Week 2008 conference

  > London, May 19-24
  > San Francisco, June 16-21
  > Melbourne, July 21-26

In-depth training:
  > 3-day Intensive Camp: Usability in Practice
  > 3-day Immersion: Tog on Interaction Design

Specialized full-day tutorials:

  > Fundamental Guidelines for Web Usability
  > Application Usability (2 days)
  > Intranet Usability (2 days)
  > Websites That Sell
  > E-Mail Newsletter Usability
  > Writing for the Web
  > B2B

Full program:
  > http://www.nngroup.com/events

DEADLINES this Friday, May 16 to:
  * Save 10% on San Francisco registration
  * Save £150 on London registration

----------------------------------

RETURNING USERS BEST PREDICTOR OF OFFLINE BUSINESS	

I gave a talk at the eMetrics Summit in San Francisco last week, having
the unenviable job of persuading a roomful of quants to try qualitative
user testing.

A major theme of this conference was the need to integrate information
about users' behavior on the website with their behavior in the physical
world.

One of the interesting presentations was by Tim Goudie (Group Manager,
Interactive Marketing, The Coca-Cola Company) who mentioned that returning
visitors were the best predictor of real-world sales. Despite this
finding, several other presenters still focused on "unique visitors" as
the key metric for their sites.

Unique visitors must be eradicated as a goal: for the vast majority of
sites, it's no good to have people visit your site once, never to return.
Uniques minus returns equals one-time tourists, which is a *negative*
metric. You're paying for traffic that doesn't engage with your brand and
is highly unlikely to be true leads or convert into customers.

I wrote about 3 types of site tourists 10 years ago:

  > http://www.useit.com/alertbox/touristvalue.html

Several speakers confirmed that Digg and other "social" services were
particularly horrible sources of tourist traffic, referring almost no
valuable leads.

Jeffrey Graham from the New York Times reported that they are able to
predict the increase in newsstand sales when big stories happen: There is
a very tight regression line that predicts next-day hardcopy sales based
on the increase in website traffic during the afternoon.

The print run order must be placed by 5PM, and if they print too few
copies, they lose sales the next day. On the other hand, printing too many
copies is very costly. The new statistics (which are very easy to compute)
resulted in about $2 M extra income for the newspaper per year.

While on the topic of Web analytics: I wrote about the relationship
between A/B testing and usability a few years ago:

  > http://www.useit.com/alertbox/20050815.html

----------------------------------

BABBAGE'S DIFFERENCE ENGINE HAD A HUMAN FACTORS PROBLEM	

This weekend, I went to see the reconstructed model of Charles Babbage's
Difference Engine at the Computer History Museum (Mountain View, CA).
There's another copy on exhibit in the Science Museum in London. Well
worth seeing a functioning version of this machine that was designed in
1850, but not built because of financial constraints.

Building the machine according to the original specifications resulted in
a working computer - with one exception: the Difference Engine runs on
human power: a person turns a crank. Unfortunately, it's beyond the
strength of any person to actually turn that crank, given the many wheels
and other parts that it drives. The modern builders had to install a
gearing unit to make it possible for a human to supply power to the
Difference Engine. (And even so, it was a pretty big guy who ran the
machine in the exhibit.)

I conclude that Babbage's machine, which is often called the world's first
computer, had a human factors problem in its original design. Luckily it
was easy to fix with some real-world testing - more than 150 years after
it was designed.

----------------------------------

FACEBOOK CLICKTHROUGH RATE 0.02%

Clickthrough rates (CTR) for Web ads have declined dramatically since the
2% levels seen 13 years ago (and still common for search ads). Now, we're
around 0.1% for most non-search ads. But Bob Gilbreath, Chief Marketing
Strategist at Bridge Worldwide reports that his Facebook ads performed
FIVE TIMES WORSE than normal ads, getting 0.02% CTR:

  > http://tinyurl.com/634ly5

Why? When visiting Facebook, people are even more focused on their task
(checking up on their friends) than they are when scanning articles on
content sites.

Banner blindness is well-documented:
 
  > http://www.useit.com/alertbox/banner-blindness.html

It's consistent with the theory behind banner blindness that it should be
more pronounced, the more the user is engaged with the website.
Thus, the more engaging the site, the less it's paradoxically worth as an
advertising vehicle.