It was recently brought to my attention how blaringly different three data sources were in reporting the "same" metric.  At the end of their email they made a random reference to taking off to get some pictures.  This random comment spurred my response:

"Comparing these three numbers is like looking at three pictures of the same thing taken with three different cameras.  Looking at the pictures, you may be able to guess what makes them different, but until you look at how the cameras that took them work, you won't know for sure why they differ.  And of course, the more experienced you are with photography, the better you are at guessing and comparing cameras."

Setting up the continuous improvement model requires the definition of three to five website goals and associated KPIs.  Now, "The wedding's over!"

There are two schools of thought - the continuous improvement process and setting targets.  I hold fast to the continuous improvement process and flee target-setting.  What?!?!  Reason #1 - Targets are for determining bonuses and if it ain't gonna determine your bonus, then don't set a target.  Real reason #1 - You just married your Goal-KPIs.  Do you really think you know them well enough to determine what are acceptable targets?  "See, we have industry benchmarks."  Phooey!  Don't strive to be like or better than your competitors, strive to be continually better than you were yesterday.  You'll fly past your competitors.  Contentment comes when you're not worried about the Jones'.

Now I'm not saying that you won't reach the point that you know your Goal-KPIs well enough to determine ample growth.  If you ever wonder if you've reached that point, then the answer is, "NO!"

So when asked the question, "How much should this KPI improve?"  Answer, "It should improve!"

On more than one occasion I have been asked, "What reports can I get?"  This question generally comes at the genesis of website tracking endeavors.  I have devised an illustrative answer to this question.  Enter "Web Analytics and Fishing - An Allegory"...

One bright and sunny day, a local fisherman was walking along the lake, tackle in hand.  A friend saw him from a distance and approached him.  He said, "My family would like fish for dinner tonight.  Can you please catch us some fish?"  So, of course, the fisherman heartily agreed.  The friend asked to see the fisherman's tackle box to select some lures and bait for the fisherman to use.  The fisherman halted his friend, "Wait, what kind of fish would you like?"  You see, the local angler knew what tackle to use for the various fish available in the lake.  "You certainly don't want carp and you may not be fond of catfish.  If I know what you want I can select the right bait and location.  You might even hold off on your dinner if you find out that what you want isn't available in the lake right now."  His friend replied, "Wow!  You are an adept angler.  How about some trout?"  The fisherman prepared his tackle and cast the bait into the perfect spot just around the bend.  In little time, the fisherman had stocked up his friend with a fresh catch of trout.  "Now wait," said the fisherman, "let me prepare them for cooking."  A little while longer and the angler handed the trout, perfectly filleted, to his friend.  As his family sat down to dine that evening, the angler's friend began, "Let me tell you about how I got these fish..."

...and so it goes with web analytics.  Provide website goals and requirements for your web analyst, allowing him/her to provide actionable solutions and analysis propelling your endeavors beyond success.

...In web analytics at least.  How many times have I heard hits, page views, visits, and unique visitors used interchangeably in emails, in meetings, and even worse - in presentations?  One of my responsibilities is to help those around me understand the terminology of web analytics.  Of course we can't make everyone in our organization read Eric Peterson's Web Analytics Demystified, but here are some suggestions for improving understanding around you:

• Create an internal webpage of web analytics terms and definitions as it applies to your web analytics implementation.  Include a link to it along with reports and presentations.  (I haven't yet done this!)
• Don't be afraid to correct someone.  Eric Peterson's "pyramid model of web analytics data" provides a great visual explanation.  (Web Analytics Demystified, p. 55)  It's not a bad idea to carry around a copy of this page in your notebook.
• Always use the correct terminology yourself and define it frequently as you see necessary.  Going back to my blog title, don't interchange terms that mean the same thing (i.e. - page views and page hits).  Pick one you like and use it consistently.
  

Many web analysts in large corporations support many teams made up of many website contributors.  Supporting this many stakeholders can be quite a daunting challenge, but with the proper organizational structure many headaches may be avoided.  The first solution one would propose is to hire more web analysts.  If you can do it, go for it!  Usually this is not the case, so here is my proposition for creating an organizational web analytics request structure that optimizes work load.

Below is a crude graphic representation of what a typical web analytics request structure should look like:

Web Analyst < Manager < Team Leader < (Team Members)
                                            < Team Leader < (Team Members)
                                                               ....
                                            < Team Leader < (Team Members)

The most important person in this request structure is the web analyst's manager, because he or she must field and clear all web analytics requests from team leaders.  The team leaders are the primary stakeholders and must field and clear requests from team members to be submitted to the manager.

This solves issues that usually result in busy work on the web analyst's part:

• Many times, a team member is looking for information that their team leader may already have.
• Sometimes Team A needs information regarding Team B's website.  Having Team A's leader request the information from Team B's leader will ensure that extra work is not done on the web analyst's part.
• Some requests are not feasible given business strategy and the web analyst's work load, therefore it may take the influence of a manager to screen such requests.