After having developed (with the help of numerous blogs and books) a my web analytics strategy over six months ago, I am faced with a new challenge - the reality that a one-size-fits-all strategy doesn't always fit.  The "bodies" are not always in "alignment" in order for a goal-based strategy to operate.  And sometimes, pages/sites need intensive analysis before a goal-based strategy can be valuable.

Now I will contradict myself.  A one-size-fits-all strategy doesn't always fit unless it is a living, breathing organism that you are willing to restructure as you go along.  By adding a new "step" to my current strategy process, it will now fit all of my current challenges, thus making it a one-size-fits-all solution once again.

So what's this "new" step?  One-time, thorough, intensive analysis; or, for short, intensive analysis.  I have placed this optional step at the beginning and end of my current process:

Ok, great, so what is inside this new box?  I put together a preliminary draft of what I'm including in my intensive analysis.  Of course, once I start using it, I expect the list and it's order to change drastically as I'm sure I will realize some obvious territory that I neglected to include.  Here is my intensive analysis list with some scribbled notes on the content of each:

1. Visitor Survey Analysis - Numbers can be picked apart, but when the customers speak, they're clear.  I pattern my surveys after the 4Q survey (Thanks, Avinash Kaushik).  It's important to get your survey up right away.  You don't want all other areas of analysis to be over and then have to wait weeks for ample responses to be collected.  Try connecting visitor feedback to metrics for stronger recommendations.  Multiple recommendations may come out of this analysis.
   
2. Heuristic Analysis - This basically means to learn about the page/site from the creator and the visitors perspective.  One area I recommend looking into is page size compared to visitor browser size and desktop size.  Another idea is to gather a list of best practices and compare the site to them.  You may not come up with any recommendations with this analysis.  The main point is for you to wrap your head around the intended purpose and architecture of the site and what the visitor sees.  It's a learning experience to prepare you for the remainder of your analysis.  Most Importantly, many ideas will pop into your head of potential areas of analysis you should perform.  Jot these ideas down and see if they fit into any of the following analysis categories.
   
3. Create Visitor Profiles - I thought of this one while halfway through this list.  Try to understand the different types of visitors to your site and create a sample profile of each type.  Keep in mind that this is just another way of being in your visitors' shoes.  Eric Peterson's book, Web Analytics Demystified, has a short section on this topic that provides a nice introduction to the topic.
   
4. Conversion Analysis - Determine what conversion(s) this page/site is concerned with and analyze the performance.  Study the conversion path and make recommendations accordingly.  There is a plethora of information available about conversion analysis.  Refer to it as necessary and then some!
   
5. Revenue Analysis - Determine (or approximate) the revenue stream of this page/site.  Approximate?  Yep, some sites are strictly for the sake of providing content (as with many of the sites I support).  However, there is usually a way to connect them to revenue even if it isn't one of their goals.  As I always tell them, "You're gonna be asked about it sooner or later; and if not, then you'll just make yourself look even better."  It's a win-win!  Execs understand dollar signs (even though they may talk about volume all day long)  ;)
   
6. Incoming Traffic Distribution Analysis - Where does the traffic originate from?  Look for areas of improvement.  Save your digging around in search traffic for the next analysis.  Segment your various reports for typed/bookmarked, website, or search engine entries.  If you want to get really crazy segment various reports for specific websites or search engines.
   
7. Search Engine Analysis - Put on your SEO hat!  Every web analyst has a little SEO in him/her.  Do I even need to detail this?  Nah!
   
8. Content Usage Distribution Analysis - This is something I got from Avinash.  How much of each content type do you have versus which types of content are your visitors spending the most time?  A combined column distribution graph will tell a ton about what you're not providing enough of (or too much of).
   
9. Navigation Analysis - Where are visitors going on your site and how are they getting there?  Is your navigation confusing?  You could spend a week on this!
   
10. Competitive Analysis - This is probably the least plundered area of web analytics.  Just looking at what all the web analytics blogs talk about reveals how little this area has been studied.  We all know about the tools out there, we just need a little nudge to start learning how to apply them (myself included).  This is last because there's no sense in studying the competition if you don't understand yourself.  How does the competition do things?  Can you improve on an idea of theirs?  Be sure to provide an overall idea of how your website is doing compared to the competition.  Everyone has a competitive streak.  :)
    

Many of these items overlap.  You'll have to decide to dig in now or write it down and save it for later.  In nearly all of these steps, you should end up with at least one recommendation.  If you have multiple recommendations, order them by considering importance, ease and ROI.  In the end, provide a complete list of recommendations in the order you think they should be considered.  You should come up with nine or more recommendations.  This intensive analysis could be done in as little as a week (if you have nothing else to do or you aren't "intensive").  Realistically, I expect that these will take me two to four weeks.

I'd love to hear feedback about other areas I missed or if you've put together such a list yourself.

For quite a while I have been circling a conundrum regarding one of my team's blogs.  The RSS XML file exceeds FeedBurners limit of 512k.  Unfortunately, I do not have access to modify the blogging tools RSS creation utility.

The quick and dirty solution:

1. Create a new RSS XML file hosted elsewhere in which the content is a summary with a "read more" link.  The easiest way to do this is to copy the existing XML and edit it appropriately.
   
2. Now edit your RSS link and redirect your feedreaders to this new file.  In my case, I only have access to change the RSS link, but I can't implement redirects  :(
   

The unfortunate reality is that, unless you have access to modify your blogging tool, you will have to manually update this every time you create a new entry.  This isn't a big deal if you don't blog frequently (as was the case with the blog in question, and, as a matter of fact, my blog too!).

Another solution I toyed with was creating a new blog in which the entries were summaries with links to the original entries on the original blog.  This would take just about as much work, but being the cautious one, I was unsure of the effect on search indexing.  I figured it might not be a big deal since the main content wouldn't be duplicated and it would provide another internal link to the content which might actually benefit search indexing.  Definitely not a stunt I need to attempt.

A good web analytics tool will offer date range comparisons such as week to week or month to month.  One of the frustrating caveats with month to month comparison is that consecutive months have a different number of days.  One way of resolving this is to normalize the data in one of the months with a mathematical equation such as:

    (total page views) +/- (difference in days) * (total page views / days in month)

Let's assume that your web analytics vendor was savvy enough to provide date range comparisons, but doesn't do such normalization for you on the fly.  Assuming your report shows the percent differences between months, here's a simple way to normalize the data on the fly.  You may even stop using the fancy normalization feature provided by your web analytics vendor!

If the difference in days is one day (for example, March to April), then there is a 3.3% difference in days.  In other words, any percent difference over +/-3.3% is more likely to be a true difference month over month.  For a two day difference, it's just double (+/-6.6%).  And for a three day difference, it's triple (+/-9.9%).

So I've been using the new Omniture Suite for about a week.  I haven't been won back yet.

 More belly-aches:

• Setting up date range comparison reports is still confusing and utterly cumbersome.
• You can't use the previous/next page table from page summary reports in dashboards.  I made the mistake of updating an Omniture 13.5 dashboard to Omniture Suite which previously had a handful of these reportlets.  I spent a good portion of an afternoon rebuilding this report in Omniture 13.5.
• Although I thoroughly enjoy the new way of moving reportlets within dashboards, I would love to see this idea combined with the "small view" that the edit dashboard feature used to bring up.
• Why aren't date range comparison reports set to sort by the most recent date range by default?
• I'm not convinced that the new email windows are easier to use.
• I found out that my looooong list of dashboards sometimes dissappears beneath the footer on short pages and reappears below it.
• It would be great to customize how many and what report suites appear in the report suite drop-down.
• It seems that I don't have campaign administration access like I do in Omniture 13.5.
• Omniture tries very hard to get their new releases out in time for their summits, but it would serve their customers better if they waited or at least stuck the "beta" tag on the release.

I know of many Omniture users who have decided to stick with Omniture 13.5 until the bugs are worked out.  It looks like I'll be joining their camp.  If Omniture gets these issues resolved they'll be back on track to winning me back, but until then it's back to Omniture 13.5 with me.

Today I finally decided to make the upgrade from Omniture 13.5 to Omniture 14, AKA Omniture Suite.  I have to say, they may have won me back.  For a while there I was getting quite annoyed with them being called the leading web analytics vendor when they were taking soooo long to add much needed web 2.0 features (not to mention statistically correct graphs).

Here are some of the wonderful things I've found in my first day in the Omniture Suite, specifically in SiteCatalyst:

• Bar graphs!  No more explaining line graphs for monthly data.
• The date range selector now shows three months instead of one.
• Twelve month calendar view.
• Drag-n-drop reportlets quickly.
• Being able to update dashboards to the Omniture Suite one at a time.
• Type and search feature for finding report suites.  For a while there I didn't realize I could type in that area and thought I had lost access to many of my report suites.
  
• My lengthy list of dashboards doesn't disappear when my mouse moves away from the list.
• I like the dashboards and bookmarks at the top, but I would rather the bookmark folders be in a drop down list similar to before.
  

Of course, there are some things that befuddle me (hopefully just due to the learning curve):

• Changing the date range resets the graph view to by day.
• The calendar is a bit slow when scrolling through months.
• The date range selector is confusing on dashboards.  I still can't figure out how to display the dashboard from a specific date like I used to be able to.  Maybe it's that whole "You may need to reformat you dashboards after updating" bug.
• No feature to update all dashboards to the Omniture Suite at once.
  
• Dashboard reportlets setup for last month still don't say what month it was.  This is annoying when you send out automated dashboards and recipients ask you what month last month is in the report.
  

After having spent the day compiling these lists while I worked, I'm beginning to wonder if they really have won me back.  There's always room for improvement.  I'll have to update my feelings over the course of the next few months with the new Omniture Suite.