More metrics

I’ll point out that trying to increase the time on site may well be at odds with usability
goals, which often are to *decrease* the time that it takes for
visitors to do things.

This obviously applies more to task-based scenarios than more general
advocacy one, but I’ll point out that the vast majority of people come
to most sites with a goal in mind – to find a piece of information, to
make a decision, take an action, etc. Facilitating their goals and
allowing them to quickly do what they’re hoping to do is a good thing.
Having so much great information that they’re compelled to stay and
read can be useful for the right sites, but I certainly don’t think it
should be the key goal of every site.

Overall, most user experience professionals dislike this concept of
“stickiness” – it implies that instead of focusing on what the end
user wants to do, we should try to ensnare them in things that will
keep them on the site. If it’s well done on the right site, a focus
on time-on- site can it can lead to really engaging stuff. But too
strong a focus can lead to clutter in too much cross-promotion,
content that isn’t useful, and if taken too far, user-hostile tricks
that literally make it take longer to do things, hard to leave the
site, etc.

So, basically, like everything else, it depends on what your goals
are. Though, of course, with Neilsen now picking up the banner, it
makes it harder to ignore.

My $0.02


Laura S. Quinn



“Google will see a drop to fifth in time spent, simply because its search
engine is intended to guide users elsewhere as quickly as possible. It
ranked third in pageviews.”

I think even the commercial folks realize that vist-length isn’t something
that can be compared site to site by iteslf. I don’t think anyone would
think that you should be penalized for having a rocking IA that speeds your
visitors to their content immediately – in fact, a decrease in visit length
may be a success metric for a redesign.

And wWho is doing the measurement, and what is their agenda? Keep in mind
that Nielsen is a business born in the era of 3-network TV, designed to
measure “eyeballs” as inventory for advertisers. As they have migrated
to the web and become a major player, they have a natural proclivity to
shift metrics to emphasize participation in rich-media content which
aligns with their TV business. Also, they have a site measurement
product to sell and they are clearly orienting it to sites that feature
rich-media content, thus helping those sites prove to advertisers that
they have an audience and therefore merit advertising money even if
traditional page-view metrics might indicate otherwise.


Measuring site success

There is no mean average for time spent on a campaign style site. Just as
with every other form of outreach, your audience will have different
thresholds for what they want to do based on the actual features you offer
and their personal interest in the campaign. The key things to measure are
not really time spent on a site overall, but actions performed working with
specific features, recidivism, and conversion rates (or achievement of goals
for all the people working with Google Metrics).

Some of the features that generally increase the time a user will spend on a
site are social networking components, blogs, forums, etc. where they can
share their thoughts as part of a community of users and build affinities of
interests with other members. The average time a user will spend
participating is influenced by a number of factors, chief among them the
level of participation by other members of the community. If there are
robust conversations and real opportunities to grow relationships, users
tend to spend more time examining what is going on and finding ways to

Recidivism speaks to the tendency of users to return to a site, independent
of any specific features, and generally web metrics tools are the best way
to track these numbers. One of the factors that chiefly influences this
metric is the use of push technology such as mailing lists, SMS components
and other forms of notification (such as alerts when new content has been
posted). It is often useful to track people who return as anonymous users
separately from those who enter the site as registered users, as the
behaviors of these groups almost universally differ.

Conversion rates speak to the tendency of users to perform specific actions.
In the context of a campaign style site, this can include participating in
an online petition, making a donation, sending a letter to an editor,
calling their representative, and other forms of action akin to those of
traditional online communities like creating user accounts, posting to a
forum, etc. Conversion rates are a key metric in that they track how
effective the online tools are that you are actually putting out there.

Creating a metric system to judge the effectiveness of all of these
components is a matter of tying numbers in each of these areas to specific
goals (i.e. do you want someone to do something, participate in something,
build relationships, etc). Time spent using specific features can be a valid
metric, but it alone is rarely a strong indicator of the effectiveness of
any particular form of outreach.

At my company, we often track specific outcomes in relation to features
using web based metric tools and other methods of analysis specific to the
tools we work with. In general, we find that 4 – 8% participation rates in
any specific category are general indicators of success at the initial
launch of a project, and look to outliers as indicators of the usefulness of
particular features. If no one is using the forums on a particular site but
10% of users are checking out the social networking features, that is
generally a good indicator that putting more emphasis on personalizing a
site will lead to better outcomes. Similarly, if 50% of users are reading
position papers but only 3% are looking into the blogs, it may mean blog
features need to be promoted in new ways to make them effective.

Something to remember is, over time, the majority of traffic to most
publicly available sites will be anonymous web users who will visit the site
once and never come back. This figure is regardless of the actual features /
options / goals that are offered, and it heavily influences the average
amount of time spent on a site. Means and modes in terms of time spent are
usually more interesting to understand; even if 90% of people are there for
under 30 seconds, the other 10% becomes the important group to target.

Michael Haggerty



Functionality and usefulness are far more important to the
success of your website than how nice and elegant it looks.

The first time I saw the Grand Canyon was a truly memorable
experience. The depth, distance and hazy rainbow of colors were
like nothing I had ever seen before. The great Colorado River
looked shoe-lace-wide down below.

We spent a day driving along the Grand Canyon and then up into
the equally magnificent scenery of Utah. But along with the
otherworldly beauty what also struck me was the poverty that
surrounded the Canyon.

For all its stunning beauty, The Grand Canyon would not be a
great place to live. Certainly, you would have a hard life if
you lived in the middle of the Canyon itself. And, given the
steepness and inaccessibility it would be hard to imagine how a
city the size of New York could develop there.

The things we think are the most beautiful are often the least
useful in a practical and functional sense. Mount Everest is
beautiful. Gold, jewellery and diamond rings are beautiful. Do
certain things increase in beauty as they lose practical

There is no question that certain designs can be made both
beautiful and functional. But for other design challenges, the
more beautiful the design is made, the less functional and easy
to use it becomes. This is particularly true for websites.

Ryanair, eBay, Amazon, Google, Craig’s List, My Space, and
YouTube are ugly websites. They are also hugely successful
websites. When I show audiences the Ryanair website, there are
audible gasps. I see people recoil from its sheer ugliness. Yet
last year, Ryanair flew 42 million passengers, and the vast
majority of them booked their flights through

Have you noticed that the Web has started to grey? There is a
severe outbreak of grey text syndrome, particularly in blogs.
Web design is falling into the trap of caring more about how a
page looks than how it reads.

Few would dispute that it is harder to read text on a screen
than in print. Most would agree that black text on a slightly
off-white background is easiest to read. It could also be argued
that font size for webpages should be slightly larger than font
sizes chosen for print.

So, why do an increasing number of websites today use small font
sizes and grey text? The answer is simple: small fonts and grew
text look better. They blend into the overall design of the
page. They are more elegant and visually appealing.

The problem with larger font sizes and black text is that they
stand out. They can dominate the page. This is exactly what
makes them easier to read. Black text in a large font stands out
from its background.

When I ask people to look at a website like Ryanair their
instinctive reaction is often to say that it is ugly. If you ask
most people to look at the most successful websites, they would
also probably tell you that they look ugly.

The fact is we don’t spend our time looking at websites. We
spend our time reading and using them. There are three things a
great web design must be: useful, useful and useful.



Most of us use Technorati to find out how popular we are, of course.  If
your ranking is high enough, you get to sit at the cool kids’ table in the

Um, yeah.  Technorati’s an interesting and useful tool, but I wouldn’t take
the exact numbers too seriously.  Theoretically, it’s tracking your blog’s
“influence” as measured by how many people are linking to you.  The
critical bits are the number of incoming links from other blogs, your
site’s “rank” (which shows the number of blogs that have more incoming
links than you over the last 180 days), and your “authority.”  “Authority”
seems to derive  from the number of incoming links, again over the last 180
days, and from the variety of sites those links come from (i.e., 100 links
from 50 sites seem to count more than 100 links from 1 site, though I don’t
know if links from more influential sites matter more than links from less
influential ones). Technorati also shows the actual sites that have linked
to you.

Technorati’s numbers fluctuate a lot (their system seems unusually buggy),
so I generally look at overall trends without focusing so much on the
details.  Main question: is my stuff interesting enough that other people
link to it, and if not, how can I improve it?  What kinds of articles tend
to draw the most incoming links and from what kinds of sites?  What other
sites are in my idea space, and what are they writing about?  The answers
to these questions can help you figure out how to build your audience and
promote your articles.

There are a ton of tips out there about building blog traffic, but the best
advice seems to come down to this: define your niche, post good content,
post frequently, pitch your articles to other sites if you think the
authors might be interested (but don’t overdo it), and look for other sites
that might run your articles with a link back to your main site.  I wrote a
few recommendations up last November, if you’re interested:

Oh yeah, that’s another one: occasionally post your articles to online
discussions, if it’s appropriate to the topic and adds value to the
conversation — but don’t do it so often that people make fun of you.


Avoiding the form bots

This kept our email addresses clean, but we did have a problem at one point.
SPAM bots kept going after one of our forms. We didn’t want to burden our
users with those PITA hoops of typing the barely legible string of
characters in the graphic tricks, so we tried something different. We have
two hidden fields on every form, and they all have a set value by default. A
real user with a browser cannot see them or change them unless they’re using
something like Tamper Data. So, the value stays the same. Bots see the
fields and change them. Our servers check the value of the fields, and if
they have changed, it doesn’t deliver the email. That has eliminated 99% of
the spam through the website.

input type=”hidden” name=”25e11d7f” value=”9e5510f” /
input type=”hidden” name=”103228ec” value=”2209ee31″ /

input type=”hidden” name=”3b8f8082″ value=”2148a195″ /

I’m assuming that the CFM file the form is submitted to compares these name/value pairs to ones it has set in the program?


Tracking document downloads in Google Analytics

So I’ve taken the code and modified it a little to make it just for
file downloads, and to cover more file extensions that the original.
This will track the following file types:

  • PDF (.pdf)
  • Microsoft Word (.doc)
  • Microsoft Excel (.xls)
  • Microsoft PowerPoint (.ppt)
  • Microsoft Visio (.vsd)
  • Microsoft Visio XML (.vxd)
  • ZIP Archive(.zip)
  • RAR Archive (.rar)
  • Text file (.txt)
  • Downloadable Javascript file (.js)
  • Downloadable CSS file (.css)
  • Executable (.exe)
  • NEW Windows Media Audio (.wma)
  • NEW Quicktime (.mov)
  • NEW Audio Video Interleave (.avi)
  • NEW Windows Media Video (.wmv)
  • NEW Mpeg Layer-3 (.mp3)

The result – some unobtrusive javascript, to use simply download the js file,
place it somewhere on your webserver and paste the following code
directly above your standard Google Analytics tracking code (which by
the way, should be as near the end of the <body> tag as possible).

<script src=”/js/taglinks.js” type=”text/javascript”></script>

(Don’t forget to replace the ‘/js/taglinks.js’ with the correct path to the javascript file)

Hey presto you’ve now got tracking for all downloadable files!

UPDATE: I’ve update the Javascript so it won’t
throw a wobbly if there are anchor links on the page, also some new
file types have been added (see above).

UPDATE: The script now tracks external links as well as document downloads.


Nielsen Shuts Window on Pageview Rankings

Yahoo reports
that Nielsen/NetRatings will be scrapping the pageview-based rankings
system that’s long comprised the industry’s webpage yardstick, in favor
of tracking the length of time visitors spend at sites.

The move comes as new technologies, such as online video and AJAX, chip at the relevance of pageviews.

AJAX in particular is tricky, as it automatically populates sites
with new information without a user having to refresh the screen or
pull up new pages. This technology renders the pageview yardstick
virtually worthless.

comScore Media Metrix, a Nielsen rival, recently developed a new
metric to weather changes in site value led by AJAX. “Site visits”
count the number of times a person returns to a site with a break of a
half hour or more.

“Based on everything that’s going on with the influx of Ajax and
streaming, we feel total minutes is the best gauge for site traffic,”
said director Scott Ross of Nielsen’s product marketing.