As practitioners in the web analytics field, we search for convergence among sources of data. When only one perspective or tool suggests insight into our users behavior we should be hesitant to jump to conclusions. As a best practice, we search for patterns among two or more sources, to increase confidence for a specific observation. Unfortunately different reporting mechanisms will often contradict each other, even in the case of the analytics superstar Google.

Here is an example that was recently brought to my attention. I lean heavily on Google Analytics to see what keywords bring traffic to the Iteration Station. Google webmaster Tools also offers a “Google only” referral report. When Google recently rolled out a data sharing component I hoped that the Google cloud would tighten up and I would see convergence among my reports.

Unfortunately, the keywords still don’t match up:

This first picture is a sample of keywords from last weeks Google Analytics keyword report:

During the same time frame, Google webmaster tools reported only three keyword phrases as entrance sources of traffic…

Clearly, the two do not match up. What is the cause of this disparity? First off, if Google Analytics is sending data to Webmaster Tools, that data is not being reported. I am hesitant to say the Google analytics is misreporting data because they record IP for each visit.

So, is this a lack of integration or instead, is Google Webmaster Tools purposefully misreporting data?

Here’s my speculation…

Google advises webmasters to use the top search queries and top clicked queries report in the following manner.

” If a relevant search query appears in the first list, but not in the second, you should work on making your site more compelling to users so that they click it when it appears in search results. Your page title appears in the results, so make sure it’s relevant and accurate. Google can take the text in your pages’ meta descriptions and display it in search results, so review your meta descriptions to make sure that they contain useful descriptions of your pages’ contents.”

Could it be possible that Webmaster Tools is giving site owners a gentle nudge to create or modify their existing content to fill holes in Google search results? By choosing to omit clicked keyword phrases, Google may compel webmasters to reshape their articles to related keywords showing up in the top search queries report. A clever trick if Google’s aim is to create a more well rounded index.

Please post your own explanation for this phenomenon as a comment and we can further the discussion.

 Here’s an explanation that I received this weekend at StomperNet Live from Dan Thies:

He explained that Google implements click tracking on only a small portion of search result pages it serves.  For that reason, the two reports do not match up.

So, Google should leverage it’s new data sharing plan to synch the two reports when available to reduce the confusion.  At the least, they should explain that the clickthrough report in Google Webmaster tools is far from comprehensive.

Have you ever wondered how your site was doing in comparison to other industry players? Well, now Google Analytics offers a glimpse into the KPI’s of your industry. Earlier this month Google allowed users to opt-in to a new data sharing program through their analytics account. The service was turned on sometime last week and data started appearing.

If you’d like to jump on the boat, go to you Google Analytics homepage and click the data sharing link:

Then check the box for “anonymous” sharing:

In their own words, Google Explains…

“Enable benchmarking within Google Analytics by sharing your website data in an anonymous form. Google will remove all identifiable information about your website, combine the data with hundreds of other anonymous sites in comparable industries and report aggregate trends in the benchmarking service. Google will also use this data in anonymous form to improve our products and services.”

Google uses some indexing mojo to determine your niche, then see how you fair against other industry verticals by choosing benchmark (beta) from the visitors dropdown. You can choose any category to compare you data to and Google promises more categories will become available as more people opt-in to share data.

Here’s an example of the benchmarking “time on site” report :

Hey +33.35% looks like Iteration Station is beating the pack!

The industry average is represented by the grey link that runs through the middle.

You’ll find reports for visits, bounce rate, time on site, page views, page/visit and % new visitors.

Besides being interesting, the benchmark reports can provide insight into where your site may lack. Low time on site: work on making content more compelling. Small visitor total- crank up the SEO. The value of data sharing is you gain a prism through which to view and evaluate the metrics of your site as compared to others on the net.

I would opt-in to the tool and check out how your efforts stack up. A little healthy awareness of the competition might be just what the doctor ordered.

Take a gander at the new flash supported interface:

Urchin differs from Google Analytics in that it is hosted on your own server and keeps track of log files. Google analytics does not keep track of your log files because it would take up too much server space. Even further, the savvy GA user will notice that GA only keeps track of aggregates of historical data so the chance for in-depth analysis is lost over time. Not so with Urchin, you can keep all your historical data on your own servers.
The mining of log files along with the rest of Urchin’s functionality offers incredibly mailable reporting.

Here’s a call out of the newest features taken directly from Urchin’s feature page:

• In-house Flexibility: Configure Urchin according to your specific requirements and run it as frequently as you wish on your own Linux, Windows, or FreeBSD servers.
• Pagetags or IP+User Agent: Choose which methodology works best for you. You can even have the pagetags make a call to your Google Analytics account and run both products together allowing you to audit the pre and post processed data.
• Advanced Visitor Segmentation: Cross segment visitor behavior by language, geographic location, and other factors.
• Geo-targeting: Find out where your visitors come from and which markets have the greatest profit potential.
• Funnel Optimization: Eliminate conversion bottlenecks and reduce the numbers of prospects who drift away unconverted.
• Complete Conversion Metrics: See ROI, revenue per click, average visitor value and more.
• Keyword Analysis: Compare conversion metrics across search engines and keywords.
• A/B Testing: Test banner ads, emails, and keywords and fine-tune your creative content for better results.
• Ecommerce Analytics: Trace transactions to campaigns and keywords, get loyalty and latency metrics, and see product merchandising reports.
• Search engine robots, server errors and file type reports: Get the stuff that only log data can report on.

Urchin offers real time reporting as opposed to the 6-18 hour lag experienced with Google Analytics. I find that this is crucial for reactive marketing campaigns where messaging should change according to real-time feedback.

The new urchin will require some heavy server configuration that may dissuade many current GA users. If you do have the installation know-how, take the three month free ride and enjoy what should be one of the finest analytics products to date.

You may also enjoy my post about Google Benchmarking

GA now offers its humble users the ability to compare two metrics at the same time. Essentially it’s an overlay line graph like the one below:

It’s available for many reports by clicking on the little dropdown that the red arrow points to in the picture below:

Once you’re there, click compare two metrics. Then just click a radio button in each column that corresponds to the metric you’d like to view. If you’re like me, I used to take screen shots of reports just so I could view them at the same time.

Use this report to determine if correlations exist between different metrics. For instance, does an increase in new visitors correspond with a leap in bounce rate?

You can also compare an individual segment to overall behavior such as the conversion rate for adwords visitors compared to the overall conversion rate.

One thing I would like to be able to see that is not available is comparing traffic sources. It would be truely insightful to be able to view a multi-line graph for direct vs. refering vs. search traffics over time.

Enjoy the new feature and let me know if there are some handy comparisons that you’ve been able to uncover.

In order to understand activity on public vs. private pages concerning such metrics as unique visitors, browser capability and organic referrals, we need generate reports on one group at a time.

You can accomplish this in Google Analytics by setting up custom filters that segment reporting to a defined set of pages.

*NOTE: Whenever you are modifying the way Google reports traffic, always set up a duplicate profile for your site. I like to call my duplicate example site sandbox. Apply the new filter to your duplicate profile and collect reports for a couple days. This will ensure that your filter is configured correctly before you apply it to your original site profile.(Learn how to create duplicate profiles )

Let’s focus on creating a profile that reports only on the public pages of our site.

Open up the filter manager at the bottom of the analytics settings page of Google analytics and click "create new filter":

Name your filter, choose custom filter from the drop down menu and click exclude.

Here’s where things get a little tricky…

Choose Request URI in the filter field box. Request URI is just a fancy way of saying use filter by a specific url requested by a site user.

If you only have one page in your public site, enter it as www\.examplesite\.com in the filter pattern box. We have to put \ before each period to tell Google "it’s just a period, not a regular expression"- more on this in a moment.

If you have multiple pages on the public part of your site we’ll have to use regular expressions to tell Google what to do. Don’t get scared, all those funky symbols are saying to Google is, "include all these URLs".

We’ll construct our expression like so (url1)|(url2)|(url3). The | symbols between our URLs operate as the modifier (or). Right now your probably saying "wouldn’t we want to tell Google to include url1 and url2. No, we wouldn’t. Google understands (and) to mean the request uri must be url1 and url2 at the same time and that’s impossible. I could ask you to bring me a beer and a chair but you’d be hard pressed to find a beerchair- unless of course I was refering to a bar stool . That’s why we have to use regular expressions (regex as the cool kids say) instead of creating multiple include filters for each page.

The ^ and $ symbols tell google where our url starts and stops. We don’t need to enter www.examplesite.com before each sub-domain because Google is smart enough to assume it’s there. So our complete example would look like (^url1$)|(^url2$)|(^url3$).

*NOTE: your homepage would be written as (^/$)

Regular expressions are a slippery slope so you’ll either have to take my recommendations at blind faith or learn about them yourself .

Once that’s set up, chose your test profile and click apply, then save changes. Wait a day and you should see only the designated pages showing up in your sandbox profile.

To create a private members only profile simply repeat this process but choose the exclude radio button.

Happy filtering everyone