When it comes to your experimentation program, you’re only as reliable as the data you use. When customer data is siloed and fragmented, it impacts the user experience and internal operations. You need to ensure your team is leveraging the most comprehensive visitor and event-level data to build better cross-device user experiences and create highly targeted experiments and personalization strategies for audiences.
In this session, you’ll learn:
- How to fuel better user experiences in Optimizely with centralized, comprehensive, and enriched data from Tealium
- How Tealium can stitch together anonymous and known visitor data and then leverage those audiences within Optimizely to generate more consistent cross-device experiences
- How to leverage both online and offline data to deliver a seamless user experience across devices
- How to connect Optimizely and Tealium using both client-side and server-side integrations for increased performance and real-time experiences
Smarter Experimentation with Fully Integrated Data
1.
2. Kyle Brierley - Director, Global Integrated Solutions
Special Guest: David Rose - Director, Customer Data & Analytics @ Gap Inc.
Smarter Experimentation with Fully
Integrated Data
3. Game Plan ● This Slide
● Tealium Overview
● Slightly Realistic Scenarios
● Assessing Your Experimentation
● Deep Dive
● Putting it All Together
● Customer Highlight - Gap Inc.
● Q&A
14. When your experiment has one variation and you get introduced to a data scientist at Opticon
“You know, I’m something of a scientist myself.”
15. Uninformed, single device experimentation
Cross-device experimentation
Cross-device experimentation leveraging all data
sources
Creating actionable insight from standardized data
across devices and data sources
Experimentation
Enlightenment
Matrix Based on
Science*
*This slide is made up and contains no
scientific research
17. Are You Swimming At Full Speed?
Unanimous Eye Roll @ Overuse of Swim Analogies
18. Experimentation
is easier said than done…
There are so many “good” reasons not to innovate.
Lack of a culture
of innovation
Complex and time
consuming processes…
Legacy systems
hinder innovation
Limited internal
resources…
A risk-averse
organization…
19. How well do you...
• Collect and standardize data
in real time
• Integrate data across end-
point solutions and contexts
• Translate data into actionable
insight
20. Squad Goals
● Marketing less reliant on engineering
● Data teams have clean data from all
sources
● Business/Exec have accessible and
actionable insights
● Product teams can make informed
testing decisions and rollouts
● End user gets a unified experience
across all devices and contexts
22. The Ideal Build
● Trackable/actionable goals and metrics
● Standardized collection of all available anonymous and customer data
● Incorporate meta and contextual data with behavioral data
● Recognize users across devices / contexts
● Congruent user experience across all endpoint solutions
● Client-side and server-side integrations for near real-time
● Centralized data sets, both structured and unstructured
● Team commitment to improvements and rapid adaptation
24. Example Goals and Revenue Drivers by Industry (tl;dr)
B2B
● Total Marketing Qualified Leads
● Average Lead Close Rate
● Call Center Deflection
● Close Rate Per Channel
● Paid vs Organic Lead Percentage
● Lead Close Time
● Customer Lifetime Value
Retail
● Total Revenue
● Purchase Completions
● Promotions
● Bundled Purchases
● Newsletter Sign-Ups
● Average Order Value
● Lifetime Value
Travel
● Total Revenue
● Registrations
● Average Trip Value
● Loyalty
● Lifetime Value
● Booking Upsells
Media and Publishing
● Ad Impressions
● Ad Revenue
● Newsletter Sign-Ups
● Subscriptions
● Page Views
● Time on Site
● Content consumption
25. Example Metrics by Industry (tl;dr)
B2B
● Downloads
● Request a Demo
● Signup / Register
● Watch Product Video
● Complete Demo
● Complete Lead Form
● Signed up for Webinar
● Search/Landing Page
Filter
● Newsletter Signup
● Watched Video
● Download Resource
● Abandon Form
● # of Products
Purchased
● Call Center Tickets
Created
Retail
● Complete Purchase
● Add to Cart
● Keyword Search
● Search / Landing Page
Filter
● Signed up for
Newsletter
● Abandoned Cart
● Used Coupon Code
● Registered for Profile
● Product Image View /
Video Play
● Average Order Value
● Average Quantity Per
Order
Travel
● Complete Booking
● Add to Cart
● Keyword Search
● Search / Landing Page
Filter
● Signed up for
Newsletter
● Abandoned Cart
● Used Coupon Code
● Registered for Profile
● Options Selecting
● Booking Upsell
Media and Publishing
● Register
● Subscribe
● Keyword Search
● Video Starts
● Video Completions
● Ad Impressions
● Ad Viewability
● Article Completions
● Image Gallery Views
● Social Shares
● Comments Left
● Pageviews Per Visitor
● Newsletter
Subscriptions
● Pay for Premium
Content
● Application
Installations
38. Data Warehousing (DB)
● Structured Data Only
● Batch Loading
● Needs Business Requirements
● Delivers KPIs and Reporting
● Reliability for Analysts
Data Lakes (Store)
● Semi + UnStructured Data
● Continuous Loading
● Adapt to Changing Requirements
● Broad Org. Applications
● Breadth of Data for ML
Two Approaches
44. How Do Tealium and
Optimizely Work
Together?
UNIVERSAL DATA HUB MEETS
THE EXPERIENCE OPTIMIZATION PLATFORM
45. Key Benefits
● Personalize based on 360 view
of customers
● Experiment across on devices in
real time
● Fast and easy
deployment/adjustments
● Page speed optimization and
‘flicker free’ fixes
● Data standardization and
collection via Tealium Data Layer
● Server-side event collection and
visitor profile updates
57. David Rose
Gap Inc. Director, Test & Learn COE
● Standards in Measurement
● Process Training
● Test Intake and Deployment
● Test Reporting and Analysis
● Tool Governance
59. Multiple Sources to Feed Experimentation
Online Behavior
Data layer connects to
multiple data repos
Offline Information
Known customers
can be stiched
between on/offline
Data Models
Platform analytics team
to produce predictive
models of our customer
behavior
KNOWING THE CUSTOMER
CONSTANTLY LEARNING
60. Optimizely and Tealium at Gap Inc.
Hundreds of Tests in 2018
Experimentation
● Data Layer → Analytics Tool
● Targeted Experiments
● Quick fixes to Known Audience
Personalized Site
200+ segment combos
Personalization
● AudienceStream integration
with Optimizely
Cross-channel
Behavior-Driven Marketing
Marketing
● Targeted Marketing combo of
online and offline behavior
● Customer Research
61. Bringing it Together - Promotional Sensitivity
Online / Offline
How much
discount do you
actually need to
purchase???
Collecting past
purchase data at a
customer level.
Data Model
Sensitivity Score
for each customer
to understand
distributions of
sensitivity
AudienceStream
AudienceStream
takes in the
production model.
Badges are
created with this
attribute.
Next Step
Do we
personalize an
offer for our
customers?
i.e., do we change
our messaging
based on their
anticipated need
for a discount?
Analytics /
Learning
Badges are
connected via
Tealium into
Analytics variable
Analysis of
purchase
patterns, visit
behaviors, etc
teach us about
each cohort of
customers
62. {Flip} Turning It Into Action
• Collect and standardize data in real time
• Integrate data across end-point solutions and
contexts
• Translate data into actionable insight