I’m excited Eric T. Peterson and Web Analytics Demystified, along with ComScore and Josh Chasin, figured out a way to overlay Web Analytics Demystified Visitor Engagement Score; this is big news.

I’m on my way to Emetrics Marketing Optimization Summit DC and writing this from my 3G IPhone using WordPress; I don’t have the ability to link to Eric T.Peterson’s post today, just now, but you can check it out for yourself at webanalyticsdemystified.com, if your interested.

Here’s the thing, I work with Web Analytics and ComScore Media Metrix, quite a lot, and have read Eric’s paper recently, so I have a seasoned opinion.

I think Engagement Scoring is going to be a big improvement for ComScore, though I hope it will be provided as part of ComScore’s core reports rather than being a new product that you have to pay extra for. Based on what information I can pull out of Eric’s post, today, I am inclined think the result will be the latter rather than the former.

Why?

Look, 4 of the 7 measures in his formula are used because the other 3 can’t be pulled out of ComScore’s Audience data, and of those the formula used, the base data is already present in any core report ComScore produces.

In his post, Peterson shows the caculated percentages for each engagement index, not the raw data and an example of how he applied his calculation. That’s what leads me to think the Engagement Score will be. Paid addition to standard reports once it’s offered, if it ever is offered (hope I am wrong on this one).

However, I do understand giving away everything valuable is also, not desirable. And, honestly,Engagement Scoring does add value, so thereis some justification for adding additional cost. Clearly, this is a philosophical issue, and I don’t know what the final form of this service will be, anyway.

But, I still think it would be unfair to tack on additional charges for calculations on data to arrive at an Engagement Score based on published and paid for data the subscriber already has.

Think of it this way, some Web Analytics Platforms, as of today, don’t give you “Bounce Rate” but it’s easy to calculate Bounce Rate based on a standard report, so why charge extra for it?

The 4 Engagement Percentages being presented in Eric Petetson’s post, that make up a property’s score aren’t much more complicated to calculate than Bounce rate is.

I think the Web Analytics Demystified Engagement Scoring of ComScore Reports, is a major advance for ComScore, but it needs to be free and offered as soon as possible.

As Eric Peterson noted, there are some features of Audience Measurement that confer a distinct advantage over Web Analytics, such as the Catogorization of web properties.

Peterson quickly came up with a loyal score for any ComScore site by dividing visits to a web property by the total visits to all web properties in the same category; it’s simple, but effective, and draws on data the Subscriber already has in a standard ComScore report, which means a subscriber can do the Loyalty Calculation for themselves, if they want.

I’m looking forward to speaking with Eric Peterson, and hopefully, Josh Chasin about the resarch being presented today, in person, at Emerrics Marketing Optimization Summit.

I want to add one more point, and this is directed more at ComScore; Data that a subscriber already gets in one format (or report) should be made freely available to a ComScore data subscriber in other forms, for no additional charge.

However, if data a ComScore Subscriber already gets is manipulated with additional data the Subscriber doesn’t have (hasn’t paid for) then ComScore, I feel, is well within their rights to repackage the data as a new product.

Examples where ComScore does not appear to be following my advice is their Cross Visiting reports, these are simply a recalulation of data a customer already paid for in the Core Reports.

On the other hand ComScore Segment Metrix adds considerable extra data to Core Reports, improving them; ComScore is entited to market Segment Metrix as an additional product because of all the additional information contained (unlocked, would be a better term).

So, from that point of view, Web Analytics Demystified Engagement scoring ought to be a free enhancement, a gift to ComScore Subscribers, who pay hand over fist for the ComScore offerings, in the first place.

What does Eric Peterson and his brand get out of it? More people know who he is, his brand spreads, which is, by all measures, a good thing.

And now my fingers tire and I’m almost in Washington DC now.

Hope there’s not too many typos and misspellings here, tried my best to avoid that.

Since being elected to the Board of Directors of the Web Analytics Association I've started to think about what kind of Social Media tools would benefit WAA members seeking to communicate with each other - to add value to the membership.

One example that comes to mine is what the New York times did with it's travel section on going to Paris and where to stay (I happen to be going to Paris tomorrow, BTW).

Readers were asked to contribute suggestions - to contribute content - and it was focused on a subject (which makes it a little different than a blog post with comments).

There's also a lot of content created by The Times with attractions on what to see, where to eat, where to stay and nightlife (with user ratings).  For example, Paris Nightlife Listings are listed and rated by readers - this is a level of "engagement".  Personally, I find such listings extremely useful.

I think this approach would benefit the membership of the Web Analytics Association Membership site - with the subjects being created and voted upon by members and moderated by WAA volunteers on subjects such as first/third party cookie debate, mobile technology, user engagement you name it.

This is part of my notes of how I would build Social Media tools that benefit members - and there's an older model - Webmasterworld.com - which had a part of the site with content that only paid members could see as well as an email  system (sticky mail) and a user community.

I'm also thinking that KickApps might be one way to build the interactivity of a site - I haven't yet seen anything that is prepackaged version of what the NY Times Travel Section put together for Paris but I think it's worth looking into, at some point, to see if prepackaged solutions exist (I am thinking past what Magnify.net and Ning.com already provide out of the box).

Talking about Engagment - what could be more engageing than a movie where your the Star?   Well, Verizon BroadBand has done it - and here's my Movie - which I made last night (below).

I just took a photo of myself late last night on my built in HP Camcorder on my HP Pavillon dv6000 laptop and uploaded it to The Action Hero site - plus I made some choices on the Plot (there were only a couple of choices for each step).

Still …. Hollywood - are you listening?   You want engagement - here's engagement.  Smart move - Verizon! :

Here's the link to the actual movie

Gary Angel's latest article in Adotas appeals to my sense of subtlety - it's often subtle, intangible elements that are the most influential, and yet, the least understood (which is why they're subtle in the first place):

"…This drive to explain everything with one simple rule is everywhere. If you listen to reports on the stock market, you’ll hear the same impulse to simplistic explanation (and with a much higher nonsense factor) – the market was “spooked” by some piece of news or was “reacting to comments” by someone. Maybe. Maybe not. There’s always something to explain a movement, but in a market movements will inevitably occur and some outside fact can always be found to match. Variation happens."

Now, it gets interesting - he goes after pageviews and uniques and says they're both flawed (we know) but for other reasons - they tend to reward the wrong things.

"….Qualified traffic is nearly always much more expensive than unqualified traffic. So if you are optimizing for traffic then you are always going to end up buying more and more unqualified traffic since it will give you more traffic for the buck. In PPC terms, this means you’ll end up buying words like “free stock quotes” instead of words like “portfolio management advice.”

When sites are optimized this way, it’s not unusual to see truly horrible campaign results – with the vast majority of traffic looking at only a single page on the site and spending less than 15-20 seconds (measured time) on the site. That isn’t an impression, it’s a blink."

I've seen way too many sites like that - "misoptimizations" he calls it and suggests that a better way to evaluate a visit is:

"… we are going to score each visitor based on how often he/she returns to the site and has a qualified visit plus how many relevant pages and relevant page time they consume in those visits."

That would be interesting, indeed - and take a bit of characterization to set up - but I think it wold be worth "measuring the intangibles".

It looks like Gary is adding to Eric Peterson's Engagement Methodology with his own way of scoring a visit.  I will have to read his article a couple of times because there's a lot in it.

I set up goal setting for a client in Google Analytics the other day and I'm surprised it was so easy.  In fact, I'm getting ideas of how Google Analytics Goal Setting can be used for an Engagement Score. 

Too bad GA allows only 4 goals - I wish the number of Goals was unlimited (or if I wanted more, they could charge me for it - in some cases, it might be worth having).  I just set this up over the last day.

The client is a hotel resort in Central America -

I used Source Referrals and set up the 4 Goals

The way this travel site is setup the contact form refreshes with a "Thank You" but doesn't change to another page - so you can't yet really tell if the form was just looked at, partly filled out and abandoned or totally filled out.  Let's set that score for a contact to be "5".

Someone who looks at the prices for the resort is somewhat interested - let's give each time the prices are looked at a score of "3".

Someone who views online video clips of the resort is also interested, but perhaps less committed - so lets  set a score of "2" for each time the movies are viewed.

Finally, someone who listens to music of the resort is also interested, but not very committed yet, so lets set that score to equal "1".

Pretty soon, you can take the chart above and create an "engagement score" for each source of traffic and decide which sources, by nature of engagement, might be more valuable.

Google Analytics - pretty neat for what it does. 

Wish list - Double the number of Goals can set - make it 8 instead of 4. 

Eric Peterson posted his 4th installment on Engagement Metrics  recently; I liked the post and noticed engagement is defined specific to  his blog. 

The problem is, without the right tools, the engagement metric, as defined below, is hard to measure; I'm not sure Engagement can be defined universally.

I tried, but with Google Analytics I was only able to get about 2 of the 5 measurements I needed.  Here's one of Eric's charts using Visual Workstation, part of Visual Sciences (see below).

This chart stood out because it does something I've tried to do elsewhere and have not been able to, given the web analytics tools I have to work with, define how "engaging" a given page/topic is to a target audience (this was a goal of one of my corporate clients, btw).

I'm wondering how much of Eric's formula would work if I looked at my own blog, webmetricsguru.com, and used Google Analytics.   Could I get an engagement score based on the following measurements (that Eric Peterson has defined):

1. Click-Depth Index: Percent of visitor sessions of "n" or more pages
2. Recency Index: Percent of visitor sessions occurring in the last "small n" weeks
3. Duration Index: Percent of visitor sessions of "n" or more minutes
4. Brand Index: Percent of visitor sessions originating directly or originating from search engine searches for terms like "Eric t. Peterson" and "web analytics demystified", etc.
5. Blog Index: Ratio of blog reading sessions to all sessions
6. Conversion Index: In this case, session- or order-based conversion 

I'm going to redefine a couple of Eric's points, above, as not all of them make sense for Webmetricsguru.com.   Here's my definition:

1. Click-Depth Index: Percent of visitor sessions of "1.25" or more pages
2. Recency Index: Percent of visitor sessions occurring sessions since January 1st (13 full days)
3. Duration Index: Percent of visitor sessions of "2" or more minutes
4. Brand Index: Percent of visitor sessions originating directly or originating from search engine searches for terms like "webmetricsguru" and "marshall sponder" etc.
5. Conversion Index: In this case, the number of AdSense Clicks

Ok, so I have 5 things I think I can measure - hopefully I can get what I want out of Google Analytics.  I'll base my numbers traffic to webmetricsguru for the first 13 days of January, 2007.

1.Click-Depth Index: Percent of visitor sessions of "1.25" or more pages

According to the chart above, I'm only getting a fraction of my traffic that's looked at over 1.25 pages of blog content - but I don't know the exact number/percentage, nor does exporting the data help me much as I can't tell the number of visitors who had more than 1.25 pages per visit viewed based on what I'm seeing.

Recency Index: Percent of visitor sessions occurring in since January 1st- much better (and the export helps as I can calculate a percentage here. 

• Total visits = 7970
• Return visits = 294

3.7% of my visitors returned in the since the beginning of the year if I am calculating this correctly.  But what about RSS feeds, aren't they supplying a lot of my content to be viewed by people that don't even have to come to my site since they can read the entire post in their blog reader.  I don't yet have the answer to that question. 

Actually, I have some confusion here as people who have come back over the last 13 days - it's  not  synonymous to Return Visitors, who visited my blog before, but not necessarily in the last 13 days.

Duration Index: Percent of visitor sessions of "2" or more minutes

I don't seem to be able to find this in Google Analytics, but I can't imagine GA not providing this informaiton.

I did get the info out of Sitemeter (see below).

Brand Index: Percent of visitor sessions originating directly or originating from search engine searches for terms like "webmetricsguru" and "marshall sponder" etc.

I haven't figured fully how to get the keyword data the way I'm used to - so I'll pass on this one.

Conversion Index: In this case, the number of AdSense Clicks.  That should be easy to do as it's defined in the analytics:

There were 775 clicks this month - percentage of total visits = 7970, or ~1%.

So, many of the measurments needed, I can't get from Google Analytics. 

I've been knocking my head against the wall, week after week, for the last 10 months working and reporting on Engagement Metrics, specificially for IBM's homepage.  But it's also true there's no actual metric for "Engagement". 

Eric Peterson has done more to try to put Engagement into an equation than anyone else.  It also helps Eric to have Visual Sciences to map the equations against real data, on the fly, and come up with interesting diagrams of Engagement by Content Area, which also draw on Gary Angel's observation that visitor segmentation needs to also reflect consumption of site content - which Eric tries to chart (see below).

"…..I know it's hard to see, but I simply dragged a bunch of pages, groups of pages, and content groups onto the page visualization map and told Visual Site to color the nodes by visitor engagement (the height of the bars represents the relative number of sessions to each node.) I could then select-in or select-out visitors based on their relative level of engagement to identify the special kinds of customers Gary refers to."

I think this is almost "there".

OK, let's go back to what I've come up as Engagement Metrics (leaving out the details - which I can't talk about anyway).  The Web Metrics platforms I have to work with are somewhat limited for measuring something as elusive as Engagement - IBM Surfaid and a pilot implementation of Coremetrics - the segmentation Gary talks about would be very difficult to do at this stage - but here's what I've come up with:

Engagement Metric 1 -  # of clicks on a specific landing page / # of clicks from a promo or specific page to the landing page.   Example: If there were 1556 clicks from the homepage to a landing page, and there were 1000 clicks on the links of the landing page …

    1000 clicks on landing page / 1556 clicks to that page = 64.2% Engagement Score

If the page were very interactive (ie: a lot of neat Flash animations and it's tagged so the Flash actions can be tracked - which I've often done) then we might see something like this:

   4000 clicks on landing page  / 1556 clicks to that page = 257% Engagement Score

So the Engagement Score for that "landing page" can be greater than 100%.  What's considered "Low Engagement" vs. Medium Engagement vs. High Engagement?  That's up to you.  Here's how I define it

• Low Engagement         = 0% - 50%
• Medium Engagement     = 51% 100%
• High Engagement         = > 100%

What that really means is …. on average, visitor are clicking on more than one link on the landing page.

Let's add Segmentation to the mix.

• Search Traffic to Landing Page = 20% of page traffic
• Intranet Traffic to Landing Page = 4% of page traffic
• Promo Ad Traffic to Landing Page = 59% of page traffic
• No Referral Traffic to Landing Page = 17% of page traffic

I could then do the same calculation on each of the segments above (ie)

- 55 clicks (visits) came from external search engines that generated 20 clicks on the landing page - the rest either went to another page or left the site.

Search Traffic Engagement = 20 clicks on landing page / 55 clicks from search engine  = 36% External Search Engine Engagement (with the content of the landing page).

And what you'd end up with is ….

• Search Engagement        = 36%
• Intranet Engagement       =  7%
• Promo Engagement          = 212%
• No Referral Engagement    = 44%

Guess where I'm going to spend my money?

Engagement Metric #2   Number of Return visits from a specific audience segment on day 1 / total number of visits from a specific audience segment on day 1  - Compared to Return visits on day 2 / total visits from specific audience on day 2 , etc.

Again, your web metrics package might not allow you to really segment audiences this way - but suppose it could - then you could do a calculation like this for a landing page about Online Gaming:

• 1200 visits from Online Gaming Enthusiasts to Gaming Page Promo during day 1.
• 50 visits (of the 1200) were repeat visits from Online Gaming Enthusiasts during day 1
• 2600 visits to the Online Gaming Enthusiasts to promo page on day 2
• 660 return visits to promo page from gaming enthusiasts on day 2

OK,    50/1200 =  4.1 % Engagement from target audience on  Day 1

OK,  660/2600 =  25.3% Engagement from target audience on Day 2

You would look at the change from day 1 to day 2, etc - ie:

Engagement Score increase by 600% from day 1 to day 2  - therefore Promotion is very engaging to the Target Audience. 

However, if there was little change from day 1 to day 2 - that would signal the content was not that engaging or there the level of engagement reached a plateau and maybe it's time to recycle the ad.

I haven't really dealt with the social media part of Engagement here - but I think it could be done, more or less, the same way - provided you can measure social activity and segment it (ie: how many of the same segment of subscribers came to my blog today vs yesterday, etc).

Engagement Metric # 3 - % of landing page traffic goes deeper into specific content of a site.

For example - if I'm an automotive site creating a page to highlight brake problems and I get 1200 visits to my brake parts landing page - how may of those visitors end up going to later look at other brake part pages on my site?   If 300 of the 1200 visitors end up looking at other brake part pages - that says 25% of the audience for a page or promo went further into the site for that topic - meaning they were "engaged" or highly interested.

All I'd have to do then is decide what my goals were (if I run an promo on brakes and only 15% of my page visits end up looking at other brake parts on my site …I'm not happy; but if 30% did look at other brake parts on my site - I'm very happy. 

There are other audience metrics I could get into - but it's getting late and I think I've written enough, for now.

I guess enough people wrote Eric back when he first published on engagement metrics yesterday - so today he offered a clarification.

"…..But Frank raises another really excellent point, albeit indirectly, should visitors who are actively interacting with RSS or XML-based content get a "push" in their engagement score? I mean, based on his response, Frank is doing two of the six things I have identified as "most important" on my site from a content and activity perspective (more on that in my next post), so if I could uniquely identify Frank I could vet whether his score is correct based on his description of his interaction with my site.

In reality I only provided Bill and Frank as examples to show how ultimately engagement needs to take real people's activities into consideration, at least as the metrics are being defined and worked out…"

Because we can't yet define every one's visit to a site (by who they are - the only sure way is through authentication) we can't really get an accurate "Engagement Score" for them, can we? 

And then again, a commentator, Robert, brought up another nail to the coffin of  a real engagement metric:

"…..Frank's response also points out the difference between the data and the customer behavior inferred by the data. In this case in particular, its not just intent being imputed, but Frank's attitude."

I'm still anxious to find out what Eric is going to come up with as an set of measurements I can apply, analytic platform agnostic, that will approximate an Engagement Score (when you don't know all the people visiting your site).