Comments on: Are "visits" the correct new metric to track site success?

By: raj

raj — Sat, 24 Mar 2007 18:00:33 +0000

Joanna: There is something called a “visitor trail”. Basically, it amounts to tracking the full session for a single IP address within a cluster of pageviews. It’s not entirely accurate if the IP address is shared, but using clustering theory and heuristics, you can get an approximate report of the pages a single visitor visited within a certain time frame and the order in which they did it. Except that you have to pay the price for getting an analytics package that will do this for you.

I used to write my own tracking code and store daily visit behaviour in a database. That was for employers and/or clients. I haven’t bothered with my own websites/ weblogs. The free analytics packages only give you partial results. It really depends on what you want to track for a given site.

By: Joanna Peña-Bickley

Joanna Peña-Bickley — Sat, 24 Mar 2007 14:31:43 +0000

As an interactive designer who uses flash and ajax with click tagging metrics to record one to one relations between visits and pageviews.

Pageviews are a god indicatr of interest in content - but not a good tracker of individuals who visited the site.

One problem for advertisers is that visits, unlike pageviews, don’t correspond to ad impressions.

But coupling visits with a site’s reach and the length of each stay offers a fuller picture of the site’s audience.

Is there a way to relate them back from an impression served?

By: Andy Beard

Andy Beard — Thu, 22 Mar 2007 19:11:42 +0000

At times I have actually hit 3 or 4 page views for new traffic. With subscribers you are looking at the 1 to 2 range as they only visit to comment and maybe click through to an old article.

The RSS metrics are hugely important., but what is impossible to determine is the quality of the subscribers.

By: raj

raj — Wed, 21 Mar 2007 00:52:10 +0000

As a web + RSS analytics person, a former search engine webmaster, and someone with a joint degree in comp sci/ math + stats, I’d have to say that the metric that is important to you depends on what you need to measure. Didn’t Mark Twain say, “There are three types of lies: lies, damn lies, and statistics.” Now, I took a “bird” course in university called “The Uses and Abuses of Statistics”. What I learned: use whatever metric works for you.

Visitor counts in and of themselves only give part of the picture. What is their behaviour while visiting? How often do they return? How long do they stay? What pages are they visting. What were they searching for, and did they find it?

Funny thing. Years ago, when I was data mining this kind of information from server logs, most of my bosses would look at my report and have their eyes glaze over. The average person who would use metrics wants it interpreted for them. So I would conclude things like “the pageviews and visits are down because college is out right now, during summer.” That’s more meaningful sometimes.