GA Summit 2012 Live Updates

This Live Blog has ended, but you can follow all past and present posts about Google Analytics Summits on our GA Summit specific page.

 

This is LIVE BLOG – we’re updating in real-time from the Google Analytics Summit 2012 at the Computer History Museum in Mountain View CA.

Follow us on Twitter as well @analyticspros for shorter updates throughout the day, and every day!

 

4:34 pm, Google Analytics Premium – Manav Mishra & Chip Hall

Year in Review

 

• Guiding Principles: Best in class service, simple and friction-free experience, and customer focused approach.
• Great response from clients: 88% Customer Satisfaction!
• More to come!

 

“Managing across multiple formats is the most significant digital marketing challenge. – DoubleClick” (paraphrased).

Analytics in a Platforms World

 

• Consumers are fast and fickle
• One-to-one marketing through a single tag
• Stages of tomorrow’s platform: build the components, make them work together, and create a single view with one access point and a common workflow

 

 

4:14 pm, Privacy & Security – Clancy Childs

Note: He is not a Lawyer and everyone should consult with your local counsel.

What does “Privacy” mean and who are the stakeholders? Site Owners, Site Visitors, Google Analytics, Government Regulators

Site Visitors:

Concerns: What information about me is shared? How do I control where my information is being used?

 

• User controls give internet users control over their data collection – Opt-out browser plugin, do not track
• In TOS, no Personally Identifiable Data (PII) collected by Google Analytics

 

Site Owners:

Concerns: Who controls my data? Is it protected? Who can access it?

 

• Named Logins (controlled in Admin interface)
• GA Support team (when authorized by account Admin)
• GA Engineering team (in order to provide service)
• Other Google Products (controlled via data sharing options but more granular options in development)

 

Privacy Policy Options:

 

• Cookie Policy Expiry can be re-set
• Truncate/Anonymize IP Address to drop the last octet of IP address
• Opt-in / Opt-out to tracking cookies that disable tracking for a specific web property ID

 

 

3:05 pm, Next Generation Measurement Panel

This session is a panel moderated by Justin Cutroni.  Panelists include: Matthew Butner, Maria Pergolino, Susan Etlinger.

Q: Tell us a little how the Universal Analytics will impact Attribution Modeling.

A: Starting to the platform out will improve the amount of insights that can be pulled but there’s still a ways to go. Still a bit challenge closing the gap between online and offline.

Q: What the ingredients needed to close that gap?

A: Within digital being able to say whether a user bought or not. Where and what frequency levels to customers engage. Information from 3rd parties to help match data sets in order to get awareness but at the end of the day still need to be able to communicate results to the CMO or other stakeholders.

Q: Are you seeing people build those types of models?

A: Everything doesn’t have to be solid. Marketing initiatives can be focused on brand lift or softer metrics. Every CMO is asking where do I invest my marketing budget for next year. Our role is to help make that decision better and put a process in place to help actively improve performance. Not every question can be answered by Google Analytics, but need to be able to see how users behave “in the real world” and psychographics to make strategic marketing decisions.

Q: What should people keep in mind?

A: What business you are in? The sales cycle. High consideration products vs. low consideration products. Your media mix. Where do people go when they use different social media channels and understanding the digital media behaviors. Understanding how external (seasonal) factors are going to affect your customer’s behavior in spite of how well your campaign may be planned. Using an experimental design. Segmenting your group out into meaningful groups so you can drive more valuable insights from each.

Q: What do you see happening in the next year and a half?

A: A lot of organization churn as companies try to figure out what to do with analysis. Finding the right mix of people with the right background and know how to ask the right questions around data. Evolution of the tool sets and getting people trained on the tools. As CMOs become bigger buyers than CIOs becoming more equipped to support those evolutions in the workplace.

 

1:47pm, User-Centric Analytics – Sagnik Nandy

Putting the “You” in You-sers.  Key points:

 

• Sessions do not spend money on sites, users do!
• Why is user-centric data important?
• Looking at simple session data will mislead you! Need to look at it from a user-centric perspective to see what’s really happening:
  • Move from “how many” to “who” and “what”
  • Move from conversions to “converter” vs. “non-converter”
  • Understand loyalty, retention and lifetime value
  • Understand true user flow

 

Google has taken a step back to re-look at the problem and build a new solution from scratch.  The result:

 

• Accumulate – it should be easy to gather user level information from multiple sources, get custom dimensions about your customers, upload [lots] of information to Google Analytics for dimension widening (50 million rows).
• Analyze – Make user centric analytics an ingrained approach to GA. Segmentation the post popular but it counts sessions. Unified segmentation will allow for visitor then visits. Segment visitors who came during specific date range. Sequential segments, visitors from organic that hit landing page then convert – a cohort. Get powerful user-level analysis
• Action – Better remarketing actions that are now user-centric vs. session-centric. Modify models and segments for better remarketing.

 

Focus on user behavior! More user-centric data is coming!

 

• Universal Analytics, Custom Dimensions, Dimension Widening
• User-level Segmentation, RFM Analysis, Cross Session Pathing Support

 

1:13pm, From Tunnel to Funnel the Omnichannel – Bill Kee

We’re back!  This next session looks to be another firehose of announcements!

First up, what does attribution look like in this new world?  Google is thinking more in “3D”.  Attribution:

 

• Used to be about click and conversion, the world is now more complicated
  • Online/Offline
  • Paid/Unpaid
• Now we’ll have path data: Display > Email > Search > Conversion

 

The issue is that need to see not just across digital channels, but across multiple screens and across online and in-store. This creates Attribution challenges and immense opportunities:

Opportunities:

 

• Budget allocation
• Higher ROI
• Customer insight
• Competitive advantage

 

Challenges:

 

• Lots of Data
• Complicated metrics
• Politics of ad spend management

 

The solution: we need better tools!

 

• Multichannel Funnels have given advertisers a much stronger picture of the conversion path
• Attribution Modeling will soon provide a different views for path comparison
• A new Custom Model Builder tool will let you build your own custom attribution models

 

Announcements & New Features!

 

• Attribution Modeling for All
• 90 Day Configurable Lookback Window
  • Understand longer buying cycles
  • Discover introducers early vs. converters later in the process
• Better Sampling Controls
  • Un-sampled for up to 1M conversion paths
  • Restrict sampling to individual conversion types for those 1 million, vs. for all conversion paths = more granular data
• Cost data for ROI Analysis
  • Automatic AdWords Cost
  • Import Cost Data & Apply to Attribution Models
  • Many new tools working to support cost data imports:
    • ShufflePoint
    • Next Analytics
    • GA Data Uploader
    • Analytics Canvas

 

Combining traffic sources and advertising into one section –  New Acquisitions Section

 

• Acquisitions, Behavior, Outcomes
• Channel Groupings in Acquisition: Customizable & Shareable

New Campaign Management Tool – Simply campaign management into a single campaign tag!  This will make marketing tagging SO much easier!

Solving the Attribution Modeling Problem

A key question for Attribution:  what model should we use? Self-driving models where attribution models are built around your data, or different defined models?

 

• In a data-driven model, the model is suggested by your analytics data
• It will examine the converting and non-converting path
• It will take that data and develop conversion probability then derive the attribution model
• Benefits:
  • Evidence basis to drive consensus
  • Transparent methodology
  • Accelerate the move beyond “last click” mindset

 

12:15 pm, Lunch Time!

We’re breaking for lunch here… which is welcomed because the keyboard is smoking from all the typing.  More Google Analytics updates, announcements, and awesomeness starting again about 1:30 pm PDT.  See you soon!

-Caleb and the Analytics Pros team

11:38 am, Avinash Kaushik Q&A

Q: How do we get customers to understand the value of analysis vs. just fixing code?

A: Have to be able to fix the code, get the data out and start to mine for analysis. Need to understand what the outcomes could be and have the competency to find those and show those to the customer. Provide 85% that is what the customer wants and sprinkle in 15% of your own insights, focused around outcomes! Learn how to prove that you can help customers make money. Have to make a “shift in trust”.

Q: Was there a tipping point or “Golden Ticket” that helped build trust with customers?

A: Traveled around the world fixing the data problem, then realized that once the data problem was fixed they still needed to know how to see the data. Had to help show them what “great” looks like so they could start seeing great. There are two ways to do advertising: (1) reactively create with existing customers and (2) proactively capture new customers. How?

 

• Reactively Create: TV Print Radio
• Proactively Capture: Social, Digital, Search, Display  

 

“Great analysts need to be great modelers.” – Avinash Kaushik

 

 

Q: Do we need to hire better/more analysts to get insights?

A: People are not always the answer. Using better tools and finding new and interesting ways to visualize data are a better way to add value and insights. Use tools like D3 to improve value.

 

11:03 am, Starting Again – Avinash Kaushik

Avinash Kaushik is starting the next session – it’s sure to be an exciting one!

Key points from Avinash:

 

• Tying data to business is hard – we need frameworks to do this [sidenote: check out our big post on a new “digital analytics strategy framework“]
• A good Custom Report should have just 3 metrics on it:
  • Acquisition
  • Behavior
  • Outcomes
• Or put it another way:
  • Owned channels
  • Earned channels
  • Paid channels

 

[pullquote_right]You can always count on senior level people not to actually understand what happens in the real world. –Avinash[/pullquote_right]

Attribution – WTH?!

 

• The real world is complicated to measure
• Before mapping out what everyone is doing, start with what we can/can not measure and draw a map
• Define. Accept. Embrace. Rock
• Don’t sell hype!

10:50 am, Recap of the Announcements So Far

Wow!  What a firehose of announcements from Google Analytics, and we’re not even 2 hours into the Summit!

Here are some recaps:

 

• Totally new analytics: Universal Analytics – open data model announced
• New features including a whole new data collection API called that is documented, extensible, cross-platform
• Improved capabilities for online to offline conversion measurement
• New abilities to measure across platforms and devices using the “universal analytics” and custom user ID’s
• New feature of custom dimensions announced – first-class citizens in GA (think Custom Variables, only in your GA dimensions list!)
• New feature of Data Importing at Scale announced – import any kind of data to GA at scale and merge with data collected by GA using the universal analytics customer ID from and Custom Dimensions

 

 

10:12 am, Data Import at Scale with Google Analytics Universal Analytics

Laura Holmes of Google is now providing details about data import at scale.  Three key parts for this:

 

• User ID based data – in beta
• Custom Dimensions – in development
• Dimension Widening – in development (target Q1 2013)

 

Using Dimension Widening:

 

• Define data based on a key (user ID)
• Upload data to GA using that unique key to join the data within GA using Custom Dimensions
• Use this for things like:
  • Tying better demographic data from a CRM to marketing data collected by GA
  • Segment against customers in certain audiences (based on widened dimensions + custom dimensions + universal analytics)

 

Example: Travelocity

 

• Can use Custom Dimensions to tie together a view of the travel buyer
• This unified view across devices, platforms, and over time that is customer specific and consistent

 

 

10:03 am, Conversion Attribution Across Online and Offline

Universal Analytics will allow Online and Offline conversions to tie together!

Here’s an example (demo based!) for offline conversion tied back to online:

 

• Demo will show tying of offline conversion back to online behavior for a car dealership
• User submits a lead form online for a car
• At the dealership, dealer receives lead in CRM system
• CRM integration shows the CRM ID/Web user ID
• Sales person at dealership assigns user ID to be a lead, then a closed win
• All of this is being logged into GA in the back-end using the new universal analytics protocol
• As events are logged by GA, some trigger new goals but tied back to online AND offline conversions for the same user

 

 

9:45 am, Multi-Device Experience Tracking

We’re now hearing about tracking across multiple devices seamlessly and with a common view of the customer.

The problem: measuring across platforms with distinct user ID’s per platform doesn’t work as devices have fragmented (desktop, laptop, tablet, phone).

We’re now seeing a live demo of the new user ID control for insight across marketing channels.

 

• Demo is starting…
• User receives an email from a daily deals site in the morning
• User responds to email on Tablet [android, we presume!] and goes to daily deal site, signs in
• Because of sign-in, user is given a unique ID for analytics based on their login-ID (anonymized of course)
• User likes it, but closes tab on tablet
• Google Analytics now sees on visit inside Analytics
• Later in the day, user comes back to email and clicks through to the deal site
• User clicks through to the site, and buys 3 of the deal
• User has completed purchase… now, what will Google Analytics see?
• With the oldanalytics, you would see:
  • 2 visits from 2 users, 1 with no transactions, 1 with a transaction
  • Marketing analysis is skewed because you don’t see a common view of the customer
• The newanalytics will show:
  • 2 Visits from ONE visitor, with 1 transaction attributed back to the user!
  • A complete view of the customer across his devices

 

9:30 am, Universal Analytics, Nick Mihailovski, Google

What is Universal Analytics?  Nick will tell us… stay tuned.

New measurement protocol

 

• Simple measurement protocol
• Reduced cookie size
• Fully documented
• Supported, official libraries

 

How can this work?  You can define the “user ID” manually and pass it into GA from anything.  Can you say time to measure that refrigerator and toaster?  Or, what about employee access scans?  Nick and John just demonstrated logging employee ID scan-ins using the new GA measurement protocol and GA real-time!

 

9:20 am, Holistic Optimization, Paul Muret, Google

Optimizing for the customer is critical.  Understanding customers across a fragmented digital landscape is key, yet a single view of the customer is impossible so far.

How many people use multiple devices in their digital day?  3+ (everyone, almost).  5+ (some!).

The problem with Google Analytics and this new, fragmented digital world is old technology from Google Analytics.  Who knows what “utm” in UTMA stands for?  Urchin Traffic Monitor.  This goes way back [circa 2003 for Urchin]!

BIG ANNOUNCEMENT!  Major change in Google Analytics platform – biggest core change yet in the history of GA.  It’s called “universal analytics” and provides an open API model for defining data your way, across devices and platforms.

 

9:10 am, Welcome keynote, Paul Muret, Google

It has been a big year for Google Analytics.  Launches include:

• Real time Analytics
• Social Reports
• Attibution Modeling
• Mobile App Tracking
• Remarketing with GA-based data
• Tag Manager
• And more…

And It Begins!

What happens when you put 600 Google Analytics fanatics together?  Why WiFi goes dead… 3 devices each = 1,800 devices at once = no Internet!