For Maximum User Understanding, Customize the SiteCatalyst Menu

stock-menu

Default Omniture report menu

Visits vs. Visitors vs. Unique Visitors…click-throughs, view-throughs, bounces…these concepts in digital analytics are fairly abstract, and many in business and marketing never really grasp the concepts fully.  Knowing the enormous amount of learning that needs to take place for digital success, why do we make our internal stakeholders hunt for data that’s organized by TOOL definitions, instead of by business function?

In this case, the “tool” that I’m referring to here is Omniture SiteCatalyst.  To be clear, there’s nothing excessively wrong about the default menu structure in Omniture, just that in my experience, understanding by end-users can be greatly enhanced by customizing the Omniture menu.

Simple modifications such as 1) Hiding Omniture variables and products not in use, 2) organizing reports by logical business function, and 3) placing custom reports and calculated metrics next to the standard SiteCatalyst reports will get users to making decisions with their data that much faster.

1)  Hide Omniture variables and products not being used

Do your users a favor and hide the Omniture products such as Test & Target, Survey, and Genesis if you aren’t using them.  Same thing with any custom traffic (props) and custom conversion variables (eVars) that aren’t being used.  Nothing will distract your users faster than clicking on folders with advertisements (T&T, Survey) or worse, frustrate the user by making them wonder “What data is supposed to be in this report?”

Just by hiding or disabling these empty reports and tools advertisements, you should see an increased confidence in data quality.  Or at the very least, keep the conversation from taking a detour.

2)  Organize SiteCatalyst reports by logical business function

Your internal users aren’t thinking about Omniture variable structures when they are trying to find the answer to their business questions.  So why do we keep our data artificially separated by “Custom Events”, “Custom Conversions” and “Custom Traffic”?

Worse yet, who remembers that the number of Facebook Likes can be found at “Site Metrics -> Custom Events -> Custom Events 21-30?”  And why are Facebook Likes next to “Logins”?  Does that mean Facebook Logins?  Probably not.

Wouldn’t it be better for our users to organize reports by business function, such as:

  • Financial/Purchase Metrics (Revenue, Discounts, Shipping, AOV, Units, Revenue Per Visit)
  • Usability (Browser, Percent of Page Viewed, Operating System)
  • SEO (Non-campaign visits, Referring Domains)
  • Mobile (Device, browser, resolution)
  • Site Engagement (Page Views, Internal Campaigns, Logins)
  • Site Merchandising (Products Viewed, Cart Add Ratio, Cross-Sell)
  • Social (Facebook Likes, Pinterest Pins, Visits from Social domains)
  • Paid Campaigns (Email, Paid Search, Display)
  • Traffic (Total Visits, Geosegmentation)

The list above isn’t meant to be exhaustive, or necessarily how you should organize your SiteCatalyst menus.  But for me, organizing the reports by the business function keeps my business thinking flowing, rather than trying to remember how Omniture was implemented by variable type.

3)  Place custom reports and calculated metrics next to the standard SiteCatalyst reports

This is probably more like “2b” to the above, but there’s no reason to keep custom reports and calculated metric reports segregated either.  Custom reports happen because of a specific business need, and the same thing with calculated metrics.  By placing these reports along with the out-of-the-box reports from SiteCatalyst, you take away the artificial distinction between data natively in SiteCatalyst and business-specific data populated by a web developer.

Why you wouldn’t want to customize?

Shawn makes two great points in his post about (not) customizing the SiteCatalyst menu: users require special training and menu customization isn’t scalable.

Users need special training

Users need to be trained anyway.  I don’t think either of us is suggesting moving all of the menus around after an implementation has been in place for years…but if you’re a company just starting out, why not start off customized?

Fellow Keystoner Tim Patten also commented to me via Twitter DM about power users being used to “default”, and it’s annoying have to learn a new menu when switching companies; I’m not really worried about power users, I’m thinking about the hundreds of users in thousands of organizations who can’t get beyond page views and visits.  Power users can pick up a new menu quickly, switch back to default, or use the search box.

This is very much true.  The larger the company, and the more complex and varied the tracking, inevitably menu customization isn’t particularly scalable.  This is probably an area where specific dashboards are a much better strategy than customizing the menus.

Summary

For me, one of the first things I look for when working with a company looking to get their digital analytics program off the ground is whether they’ve customized their Omniture menu structure.  As a free customization, it’s something that companies should at least consider.  Organizing reports by business function requires a business to think about the questions they want to regularly answer, will keep novice users from focusing on implementation concepts, and overall is just better because it’s how I think 🙂

This blog post is a continuation of a Twitter conversation with Shawn C. Reed (@shawncreed), Jason Egan (@jasonegan), Tim Patten (@timpatten) and others.  Shawn’s counter-argument can be found here.  Jason wrote about Omniture menu customization a few years back.  And finally, if you want to read more pros-and-cons about SiteCatalyst menu customization, see the Adobe blog posts here and here.


Effect Of Modified Bounce Rate In Google Analytics

A few months back, Justin Cutroni posted on his blog some jQuery code that modifies how Google Analytics tracks content.  Specifically, the code snippet changes how bounce rate and time on site are calculated, creates a custom variable to classify whether visitors are “Readers” vs. “Scanners” and adds some Google Analytics events to track how far down the page visitors are reading.

Given that this blog is fairly technical and specific in nature, I was interested in seeing how the standard Google Analytics metrics would change if I implemented this code and how my changes compared to Justin’s.  I’ve always suspected my bounce rate in the 80-90% range didn’t really represent whether people were finding value in my content.  The results were quite surprising to say the least!

Bounce Rate - Dropped through the floor!

bounce-rate-graph-google-analytics

Starting April 24th, Bounce Rate drops considerably!

As expected, implementing the content tracking code caused a significant drop in bounce rate, due to counting scrolling as a page “interaction” using Google Analytics events. Thus, the definition of bounce rate changed from single page view visits to visitors that don’t interact with the page by scrolling at least 150 pixels.

In the case of my blog, the bounce rate dropped from 80-90% to 5-15%!  This result tells me that people who arrive on-site aren’t arriving by accident, that they are specifically interested in the content.  Sure, I could’ve validated this using incoming search term research, but this provides a second data point.  The content I provide not only ranks well in Google, but once on-site also causes readers to want to see what the article contains.

Readers vs. Scanners

Even with the bounce rate drop above, I really don’t get a good feeling about whether people are actually reading the content.  Sure, people are scrolling 150px or more, but due to the ADHD nature of the web, plenty of people scroll without reading just to see what else is on the page!  That’s where the “Readers vs. Scanners” report comes in:

google-analytics-reader-vs-scanner

62% of visits only scan instead of read - Need to do better here!

The report above shows that only 38% of visits to the site actually READ an article, rather than just quickly scroll.  This is disappointing, but now that I’ve got the information being tracked, I can set up a goal in Google Analytics with the aim of improving the ratio of actual readers vs. quick scrollers.

Average Visit Duration - Still useless

Like the bounce rate definition change above, average visit duration and average time on page also change definitions when using the jQuery content tracking code.  Given that Google Analytics calculates time metrics by measuring the time between page views or events, by adding more events on the page, all time on site metrics have to increase (by definition).

avg-visit-duration-google-analytics

Hard to see because of the Y-axis, but Avg. Visit Duration increases significantly as well.

That said, average visit duration is still a pretty useless metric, given that an increase/decrease in this metric doesn’t immediately tell you “good” or “bad”…

Content Consumption “Funnel”

Finally, the last change that occurs when you implement the content tracking code is a series of Google Analytics events that measure how far down the page visitors are actually seeing.  This report, in combination with the Readers vs. Scanners report, helps understand reader engagement better than any generic “Time on Site” metric can do.

content-consumption-google-analytics

From this report, I can see that of the 2,102 articles loaded:

  • 89.4% of the articles have a “StartReading” event fired
  • 89.8% of those who start to read an article reach the bottom of the article.
  • 19.7% of those who reach the end of the article scroll past the comments to reach the true end of page

The first metric above is analogous to subtracting the bounce rate from 1, the percentage of articles viewed that don’t bounce.  The second metric (complete articles seen), with a success rate of 89.8% is ripe for segmentation.  I stated above that only 38% actually READ an article, so segmenting the above report by “Readers” vs. “Scanners” will surely lower the success rate in the “Readers” population.

Finally, that <20% actually touch the true bottom of page is surprising to me, since this blog really doesn’t get many comments!  If there were thousands of comments and the pages were really long, ok, no one sees the bottom…but here?  I’ll have to think about this a bit.

Great update to Google Analytics default settings!

Overall, my impression of the jQuery code snippet developed by Justin and others is that it is extremely useful in understand interaction of visitors to content sites.  The only downside I see here is that it changes the definition of bounce rate within Google Analytics, which could be confusing to others who 1) aren’t aware of the code snippet running on-site or 2) don’t quite understand the subtleties of Google Analytics implementation with respect to Events and the non-interaction setting.

But since this is my personal blog, I don’t need to worry about others mis-interpreting my Google Analytics data, so I’m going to keep this functionality installed!

Update 7/25/12:  Google Analytics published a similar method to the one described above, using “setTimeout” to modify bounce rate based solely on time-on-page.


Adobe Discover 3: First Impressions

Adobe Discover

With yesterday’s code release, Omniture Adobe released version 3 of their “Discover” tool, THE way to perform web analysis within the Adobe Digital Marketing Suite.  While SiteCatalyst has its place for basic reporting, to really dig deep into your data for actionable insights there’s no substitute to using Discover.

But as with every product overhaul, there is the potential to change things that users liked and while not make enough improvement to excite the user base…but luckily, that’s not the case with Discover 3.  Here’s how I see the new features and design changes.

New “Darth Vader” interface

adobe-discover-3-screenshot

"Ooh, tough looking. Just like hardcore web analysts!"

Of all the cool things about Discover 3, I’m not sure the new color palette is one of them.  Several reasons were given by Adobe for choosing the carbon colored interface, from trying to match analyst’s personalities (yuck!), reducing eye strain (ok), and consistent branding (eh).  Of the three, I’ll say that reducing eye strain is a worthy goal, although Discover 3 never struck me as “eye-burning” in the past.

Maybe I’ll grow to like it, but right now, it seems really dark.  The light gray text on dark gray background needs a bit more contrast, and in general, the interface feels kinda depressing.

Calendars - No more #^%&$ sliders!

Now we’re getting somewhere.  The slider interface in Discover 2 never made sense to me.  You pick your time period up front, open a report, and then to modify the time period within an individual report you needed to move a bunch of jerky sliders around.

In Discover 3, we now have the same style calendar interface as SiteCatalyst.  Makes sense from a consistency standpoint within the Adobe Digital Marketing Suite and a general UX standpoint.  Pointing at two dates on the calendar is way easier and faster than moving endpoints of a slider!

Heterogeneous Pathing

This is so completely badass and the best new feature of Discover 3.  No longer are you confined to a fallout report that only includes just one Omniture variable type.  So if I want to do a funnel that measures visits containing a few different pages, then triggering a Facebook ‘Like’ event, a Cart Open, then an Exit Link, I can now do so!

You can also switch from “Visit-level” to “Visitor-level” on the fly, which can also be useful depending on how your view your business.  Some people like to think about every visit being an opportunity to convert on-site, whereas Avinash advocates in his Web Analytics 2.0 book that using Visitors as the denominator for conversion rate is the proper thought model.  I won’t weigh in on the difference in this post, but it’s cool that we can now change back-and-forth to see what the differences in the data are.

Table Builder

adobe-discover-3-table-builder

Nice drag-and-drop options, very PivotTable like

Finally, the last really obvious difference between Discover 2 and Discover 3 is the table builder while using ranked reports.  Like the eye-strain issue talked about above, the amount of time that it took for reports to build never really seemed like an issue to me.  Perhaps that’s the SAS programmer side of me that often waits hours to return a result of a complex set of commands.

But now that I’ve used the table builder, it’s definitely an improvement on how data tables get built.  You get to specify each element you want in the table first, THEN the data gets retrieved.  It may sound like a small change, but when you already know what you want, not having to wait for the table to build while you keep dragging in metrics does feel like it’s way faster to get the table you are looking for.

Adobe Discover 3 - Definitely an improvement

There are probably 20 other things I haven’t noticed yet in the new Discover 3 interface, but from what I have used so far, this is a great upgrade in functionality!  It feels faster to get things completed with the table builder and the new pathing functionality across all variable types is a long time coming.  Now, if only there was a different color palette I could choose, it’d be perfect…maybe something like this?

omniture-discover-1.5

You should prefer green, not carbon.


  • RSiteCatalyst Version 1.4.12 (and 1.4.11) Release Notes
  • Self-Service Adobe Analytics Data Feeds!
  • RSiteCatalyst Version 1.4.10 Release Notes
  • WordPress to Jekyll: A 30x Speedup
  • Bulk Downloading Adobe Analytics Data
  • Adobe Analytics Clickstream Data Feed: Calculations and Outlier Analysis
  • Adobe: Give Credit. You DID NOT Write RSiteCatalyst.
  • RSiteCatalyst Version 1.4.8 Release Notes
  • Adobe Analytics Clickstream Data Feed: Loading To Relational Database
  • Calling RSiteCatalyst From Python
  • RSiteCatalyst Version 1.4.7 (and 1.4.6.) Release Notes
  • RSiteCatalyst Version 1.4.5 Release Notes
  • Getting Started: Adobe Analytics Clickstream Data Feed
  • RSiteCatalyst Version 1.4.4 Release Notes
  • RSiteCatalyst Version 1.4.3 Release Notes
  • RSiteCatalyst Version 1.4.2 Release Notes
  • Destroy Your Data Using Excel With This One Weird Trick!
  • RSiteCatalyst Version 1.4.1 Release Notes
  • Visualizing Website Pathing With Sankey Charts
  • Visualizing Website Structure With Network Graphs
  • RSiteCatalyst Version 1.4 Release Notes
  • Maybe I Don't Really Know R After All
  • Building JSON in R: Three Methods
  • Real-time Reporting with the Adobe Analytics API
  • RSiteCatalyst Version 1.3 Release Notes
  • Adobe Analytics Implementation Documentation in 60 Seconds
  • RSiteCatalyst Version 1.2 Release Notes
  • Clustering Search Keywords Using K-Means Clustering
  • RSiteCatalyst Version 1.1 Release Notes
  • Anomaly Detection Using The Adobe Analytics API
  • (not provided): Using R and the Google Analytics API
  • My Top 20 Least Useful Omniture Reports
  • For Maximum User Understanding, Customize the SiteCatalyst Menu
  • Effect Of Modified Bounce Rate In Google Analytics
  • Adobe Discover 3: First Impressions
  • Using Omniture SiteCatalyst Target Report To Calculate YOY growth
  • Google Analytics Individual Qualification (IQ) - Passed!
  • Google Analytics SEO reports: Not Ready For Primetime?
  • An Afternoon With Edward Tufte
  • Google Analytics Custom Variables: A Page-Level Example
  • Xchange 2011: Think Tank and Harbor Cruise
  • Google Analytics for WordPress: Two Methods
  • WordPress Stats or Google Analytics? Yes!
  • Building a Data Science Workstation (2017)
  • JuliaCon 2015: Everyday Analytics and Visualization (video)
  • Vega.jl, Rebooted
  • Sessionizing Log Data Using data.table [Follow-up #2]
  • Sessionizing Log Data Using dplyr [Follow-up]
  • Sessionizing Log Data Using SQL
  • Review: Data Science at the Command Line
  • Introducing Twitter.jl
  • Code Refactoring Using Metaprogramming
  • Evaluating BreakoutDetection
  • Creating A Stacked Bar Chart in Seaborn
  • Visualizing Analytics Languages With VennEuler.jl
  • String Interpolation for Fun and Profit
  • Using Julia As A "Glue" Language
  • Five Hard-Won Lessons Using Hive
  • Using SQL Workbench with Apache Hive
  • Getting Started With Hadoop, Final: Analysis Using Hive & Pig
  • Quickly Create Dummy Variables in a Data Frame
  • Using Amazon EC2 with IPython Notebook
  • Adding Line Numbers in IPython/Jupyter Notebooks
  • Fun With Just-In-Time Compiling: Julia, Python, R and pqR
  • Getting Started Using Hadoop, Part 4: Creating Tables With Hive
  • Tabular Data I/O in Julia
  • Hadoop Streaming with Amazon Elastic MapReduce, Python and mrjob
  • A Beginner's Look at Julia
  • Getting Started Using Hadoop, Part 3: Loading Data
  • Innovation Will Never Be At The Push Of A Button
  • Getting Started Using Hadoop, Part 2: Building a Cluster
  • Getting Started Using Hadoop, Part 1: Intro
  • Instructions for Installing & Using R on Amazon EC2
  • Video: SQL Queries in R using sqldf
  • Video: Overlay Histogram in R (Normal, Density, Another Series)
  • Video: R, RStudio, Rcmdr & rattle
  • Getting Started Using R, Part 2: Rcmdr
  • Getting Started Using R, Part 1: RStudio
  • Learning R Has Really Made Me Appreciate SAS