Re: [webanalytics] Time and Space?
- Awesome question but it depends on the situtation. Page Views are not relevant with Flash site so we need to measure, perhaps, how long the average visit session is. But let's be careful, page views and time spent per visit are quantitative concepts. The question becomes how can we take either of those metrics to gain insight into qualitative behavior towards what your visitors are doing on your site. For instance, are my visitors watching that video (such as on altrec.com) that lasts 1.5 minutes and then purchasing the product promoted on it? I hope this makes sense.
----- Original Message ----
From: josh chasin <jchasin@...>
Sent: Friday, June 29, 2007 10:03:33 AM
Subject: [webanalytics] Time and Space?
I have a question for the Web Analytics community that is not,
strictly, a web analytics question.
Advertising media is a lot like physics-- in the sense that both are
all about time and space. In the early '90s, the Internet
advertising business community adopted the construct of the page view-
- essetially deciding to align around a space-oriented model, akin to
print, as opposed to a time-oriented model, akin to broadcast.
Personally, I always thought that the space construct was not the
right way to go; people spend time with tie Internet, as opposed to
consuming pages. Certainly, the technology environment at the time
contributed to and defined the basis for the consensus.
With new technologies like AJAX, and with widgets and web aps,
perhaps the time has come to revisit time.
So this is my question. How do you folks feel about the idea of
Internet audience metrics migrating from a spatial construct (the
page) to a temporal one (time spent/duration) ? Let me remind you
that I work for comScore, so I'm thinkng largely bout advertising
metrics and reporting as opposed to WA (although obviously this would
be germane to WA.)
Yahoo! oneSearch: Finally, mobile search
that gives answers, not web links.
[Non-text portions of this message have been removed]
> Josh,I agree one challenge is that you really can't compare these metrics
> One challenge is that you really can't compare these metrics across
> different sites with consistency.
consistently across different sites.
Space and time just describe relationships between objects and never
absolute values. If you define an abstract context where all objects can
be represented in relationship with it, let's say an Euclide Geometric
Space, then you can start to look for qualitative values such as: bigger,
smaller, faster, etc...
We will be able to compare these metrics across sites when we will be able
to define an abstract context like the Euclide Geometric Space represents
for space and time.
Then all interactions can be measured in relationship with the defined
context, so we could measure qualitative values as: interest level,
loyalty, lead degree, readiness to act, and so on.
Technologies for marketing
- I saw a quote in the WebMetricsGuru blog that made me think about some
of the ways we are measuring engagement.
"The longer a purchasing decision takes, the more important a
customer's safety and trust in a Brand, Product/Service become."
If you want to see the full post from Marshall Sponder, check out
I think the basic point here is that very few web-based decisions
happen in a vaccuum. I'd like to discuss the implications for
measuring engagement. Is the point to get them to do things on the
website or is it to get them to take an action or make a decision.
This doesn't apply just to sales oriented site. I spend my days
working on a leads-oriented site where the desired decision could take
12-24 months. Consider a sequence of visits that begin with banners
linking to a landing page and a quick exit (scanning visit). Later
this visitor returns to the site and looks a little deeper. Maybe
reads articles and uses applications (high engagement?), then leaves.
Finally, after lots of discussion and/or investigation, the visitor
returns (several months later) and fills out a form requesting a
contact or more information (maybe another short visit).
Is it enough to develop engagement measures or KPIs that focus on a
single visit? How do we assess this over time? What if we can't use
persistent cookies? What if our users don't allow the cookies?
> This doesn't apply just to sales oriented site. I spend my days working onIn Marketing we would call this "sales pipeline". It allows management to
> a leads-oriented site where the desired decision could take 12-24 months.
> Consider a sequence of visits that begin with banners linking to a landing
> page and a quick exit (scanning visit). Later this visitor returns to the
> site and looks a little deeper. Maybe reads articles and uses applications
> (high engagement?), then leaves. Finally, after lots of discussion and/or
> investigation, the visitor returns (several months later) and fills out a
> form requesting a contact or more information (maybe another short visit).
forecast sales in the future. Many of the popular "CRM" apps like
salesforce.com provide this functionality, the ability to see milestones and
flag events that should have occurred but did not.
You find this same kind of utility in project management software. Us
Marketing folks would like to get to the point where we can manage customers
and segments as "projects" with tools like this...
> Is it enough to develop engagement measures or KPIs that focus on asingle visit?
I'm not sure what percentage of the folks in this forum are Marketing
people, but it seems to me - at least from the posts - most people here have
IT backgrounds and are analysts, so I'm glad to see questions like this.
The answer to your question is no, it's not enough to focus on visits.
In fact, the *lack of visits* is more important in many cases, and
especially if you are talking about "engagement". In my opinion, no visits
= not engaged.
There is certainly a need for understanding "intra-visit" engagement. But
the idea of "over time" or "days since last event" is also important from a
Marketing perspective, since even if a visitor is totally engaged in a first
visit, if they never come back for a second visit, most Marketing folks
would have a hard time calling the visitor "engaged" - or the campaign that
produced such a visitor a success.
Eric Peterson pointed this idea out in his original posts on measuring
engagement by including this idea:
"3. I would like visitors to maintain a low recency with my content,
regardless of whether they're reading blog posts or viewing pages on my
And he is right about this. Engagement as a concept must occur in two
planes to be useful to Marketing (and I would argue also to Management):
1. A Consumption plane, which is "counting" or Frequency oriented and
speaks to the actual value of activities that have taken place. How many
different activities and how often? Much has been already said about
constructing weighted content or feature averages and so forth for this.
2. The "days since" plane, which to me is the heart of engagement and the
reason you should care about this idea. Otherwise, engagement is not really
a new idea, it's just a more complex version of Content Grouping.
I would argue, you can't be really engaged in something NOW if you don't
have a likelihood of continuing on with it, can you? The state of High
Consumption in the past with infinite "days since" is not engagement, this
state is "not engaged". In this plane, you report on the "days since",
"days between" or some other time-based metric.
If you don't have both these planes, you can end up with paradox: a visitor
/ customer with low Frequency / Consumption cannot be highly engaged.
From a Marketing perspective, this is simply not true. New customers / new
users would often end up in this category - very engaged with the site, low
activity. First visit with a very specific search phrase, for example.
What does the concept of engagement mean? If it is to include a "days
since" kind of idea - and I think it should - then it must be true it is
possible to be engaged *without* current activity, right?
Then at some point, when "days since" expands to some topic-specific
threshold, the visitor / customer is "not engaged". During the time when
this "days since" window is expanding, engagement is falling, relative to
those visitors with a shorter "days since" window.
For two visitors with equal amounts of activity (Consumption), the one with
shorter "days since" is the more engaged visitor.
This is a critically important idea to Marketing, because just as in the
sales pipeline / project management examples above, one can set milestones
for "days since" and flag or trigger events based on this component.
Why is this capability important? Because now you are crossing into the
world of prediction - predicting which content, visitors, customers,
campaigns, all of it - will create the most value in the future.
For Marketing and Management in general, predictions are a lot more powerful
tool than reports on what has happened in the past.
Web Site: http://www.jimnovo.com