Over the last year, the Standards Committee have been hard at work. We took the definitions document from last year, and have totally revamped the definitions based on your feedback. We've added a few new definitions, but one of the things we're most excited about is our new "Ask Your Vendor" questions, where we've highlighted places where different vendors do calculations in different ways.

One thing we're doing differently this year is to solicit feedback via the Web. In fact, we're looking for your feedback on this very blog, where we'll be posting each definition and giving the web analytics community a place to discuss each one. We really want to know what you think: good, bad, or indifferent!

Stay tuned at this address... we'll link to each definition as we get them posted. In the meantime, you can download the document here [PDF].

My co-chair, Judith Pascual from ZAAZ, and I would like to give a huge "Thank you" to our committee members who have worked so hard on our document. We realize that they don't get nearly as much recognition as they deserve! And we want to give a special shout-out to Jodi and Jason:

Jodi McDermott from Clearspring Media -- In addition to her contributions  in our regular committee, Jodi is co-chair of the Social Media Standards subcommittee. Thanks Jodi!!
Jason Burby from ZAAZ -- My former co-chair, Jason has been an invaluable guiding force on the Standards Committee for years! Thank you Jason!

And here are the folks who made our 2008 revision happen (in no particular order):
Aseem Patel from Motley Fool
Ned Kumar from Fed Ex
Anna Long from Web Analytica
Shaina Boone from Critical Mass
Katie Cipollone from Saks 5th Avenue
Bob Russotti from ANSI
Sarah Farebrother from Gazetter Communications
Brian Perry from Unica
Akin Arikan from Unica
Tim Kraft from Yahoo
Penny Tietjen from Intuit
Scott Oliver from IBM
Josh Dreller from Fuor
Elizabeth Magill from Coremetrics
Steve Harris from Clearspring Media
Jim Cain from Sitebrand

We also have several newer committee members who have recently become active, and we are looking forward to their contributions as we assimilate public feedback on the document.

Comments

[SEO South Africa]

I am new to Analytics and believe this document may help me understand better what everything is, thank you.

[Jaume Clotet]

My comment is about the official definition of web analytics "Web Analytics is the measurement, collection, analysis and reporting of Internet data for the
purposes of understanding and optimizing Web usage." In my opinion the word "user" is missin in this definition I should suggest something like "Web Analytics is the measurement, collection, analysis and reporting of online users interaction data for the
purposes of understanding them needs and optimizing Web usage."

Thanks,

Jaume

[Akin Arikan / Unica]

So happy to see the fruits of the Standards committee out in the public and available for commentary by all & everyone.

Just noticed one thing that had escaped my attention earlier in the WAA's official definition of web analytics itself:

"Web Analytics is the measurement, collection, analysis and reporting of Internet data for the
purposes of understanding and optimizing Web usage."

I would propose to amend this definition just slightly to "Internet and related data".

Why?

Because increasingly companies are segmenting web usage by data such as customer account information, product information, etc. And in the future no doubt more companies will segment online usage by offline behavioral segments. So I think this small change is worthwhile.

Many thanks to the Standards committee for all the hard work over so many years!

Akin

[Akin Arikan / Unica]

We missed a typo with Visitor Referrer. It should be a dimension, not a count.

[Akin Arikan / Unica]

I wonder whether we missed something while defining Bounce Rate or Entry Pages. Namely, we defined Bounce Rate as:

"Single page visits divided by entry pages."

But we defined "Entry Page" as type Dimension. It should have been dimension or count in order to be suitable for a calculation, no?

[angie]

SEO South Africa:
Thanks for letting us know!

Jaume:
I appreciate your suggestion. That definition has been around for awhile, and give your feedback and Akin's, it may be time to revisit that one. You have a good point about mentioning that this isn't just any data, but data relating to users (people!).

Akin:
Regarding the definition, I agree, and we should probably go back and revise this one to reflect more modern times.

Thanks for the other two catches. Will put them on our list of things to revisit.

[Rich Page]

Great stuff!

However, regarding bounce rate, I would point out in the definition the circumstances thats its not very useful (and potentially misleading). I wrote an article about it on my blog:
http://rich-page.com/win-at-web-analytics/revealed-when-bounce-rate-sucks/

Worth a mention?

Also, regarding your 'exit page ratio' i find it more useful to use 'visits' instead of 'pageviews' in the calculation. Particularly when an exit is an exit from the session (visit). I mention this in my blog post too.

Thanks!

Rich (a web analytics fanatic)
http://www.rich-page.com

[James Schend]

I'm sure it would be too large a task for this year's document, but I wanted to ask if there were any plans to add metrics more specific than pageviews. I notice that you do include the concept of "Events" but only in a generic manner.

We (Razorfish Site Optimization) have an analytics product that tracks (virtually) every action on a page so that we can advise our clients on things such as in-form drop-off rates, data that would be difficult to get with a more conventional analytics package.

I'd recommend classifying events by the type of action taken by the user. At the most basic level, distinguishing between Loads, Clicks, Scrolls, Unloads (data which is extremely important in determining user engagement with a site; Load/Unload to give a detailed time-on-page value.)

To go a step further, you could define an event type for every event that Javascript event handlers can be installed on: Load, Click, Scroll, Resize, Focus, Blur, Change, Keydown, Unload. In addition to those, we use Mouseover, Mouseout as required.

Feel free to contact me if you have any questions or want to discuss this idea. Obviously, we already have our own terminology we use for our own projects, but it would be nice to have some standardization of Events in the interest of analytics vendors with more precise data.

[Cindy Dadon]

This is a great document, but I have several revisions to share.

Events - there should be a mention in the notes that events can also capture activities around otherwise immeasurable activities, such as clicking off the site and clicking on a "download a PDF" button. And while I understand where James from Razorfish is coming from in terms of creating a wider scope of definitions for the "event" I don't think many companies/websites are able to get to that kind of detail using their primary web analytics tool, so my suggestion is to keep this definition document basic.

Visits- there is a mention that analysts should as their WA vendor whether a "Google search>your site>yahoo search>your site" counts as a new session or an external referrer. This indicates that this type of visit definition could be vendor-specific. I would caution against using this same example as default in other definitions, such as "entry page," and "landing page."

Landing Page - there is a sentence in here that says - "A useful metric may be number of landing pages per session which indicates the degree that visitors move on and off a web site." I strongly disagree with this statement. This metric would indicate how navigable the site is, and, depending on the measure of success of the site, whether the visitors were engaged in the site's content offering.
Also, in what situations would it be "appropriate to limit landing pages to only those pages with an external referrer"?

Visit Duration - Ever since I understood how Omniture and Google Analytics sets up the algorithm to calculate "time spent", I have disqualified this metric as inaccurate. Basically this metric is calculated by averaging averages, and I find that out of all available metrics in a tool, this is the least accurate. Also, I believe that including "zero duration" visits can harm your calculated average visit duration, no? Wouldn't this skew your time spent downwards?

Referrer - There are certain types of server/client side redirects that cause the referrer string to be erased and thus a referral to look like "direct load" or "no referrer given". This was a big deal about 4 years ago when I first found out about this, and I believe this would be a valuable nugget to put in the notes.

Page Referrer - since there are no notes in this definition, maybe it would be worthwhile to mention that in a one page visit, the page referrer and the visit referrer are the same.

Click-through - typically when there are large discrepancies between the "sending side" and the "receiving side", this can indicate that the destination page may not be loading correctly, and that people are repeatedly clicking on the link but not getting a page loaded.

Return/Repeat Visitor - I have always used these two words "return" and "repeat" visitors interchangeably, and I have heard other people refer to these as one and the same. Why is there a differentiation between the two? I've always thought of the "visitor" metric as giving me the most accurate count of how many "people" came to my site, and then the qualifier "new" or "repeat" as telling me whether this person has ever set foot on my site before. As it goes, you cannot be definied as both a "new visitor" and a "repeat visitor" in a given time frame. I believe that the differentiation and examples given can be confusing to the average web analyst looking to this document for guidance.

Repeat Visitor - I believe that this is referring to "Repeat Visits". the notes section defines this metric as "the measure of the amount of visit activity during a reporting period". This sounds like "repeat visits" to me.

That's all I had.

I am very impressed by the success of this document and how it has highlighted the importance of having standard definitions in a still-blossoming industry.

Thanks,
Cindy

[Tenly]

Many Vender use Clicks instead of Hits.

Clicks is more popular than Hits.