mau vegas 2016 — conceptualizing and implementing a cross-channel segmentation strategy
TRANSCRIPT
CONCEPTUALIZING AND IMPLEMENTING A SEGMENTATION STRATEGYMAU 2016Cody Ryan – Senior Manager, User Retention @
TODAY’S CONVERSATION ON SEGMENTATION1. Why it matters2. What data you should use3. Stages of segmentation4. Results of improved segmentation at Ibotta5. Key takeaways6. Q & A
SEGMENTATION OVERVIEW
WHY IT MATTERS ALL USERS ARE DIFFERENT! Users differ and are unique based on:
Demographics In-app behavior Needs Socio-economics Beliefs/perceptions Motivations
WHAT TO SEGMENT ONBehavior
Demographic
Acquisition Source
Predictive Modeling
User-level Data Warehousing
Proper data tracking is the cornerstone to segmentation
Segmentation Strategy
High ValueLow Use
High ValueMid Use
High ValueHigh Use
Mid ValueLow Use
Mid ValueMid Use
Mid ValueHigh Use
Low ValueLow Use
Low ValueMid Use
Low ValueHigh Use
IF YOU DO IT RIGHT… …segmentation drives User Acquisition (UA)
The easiest way to drive increased value and decreased churn, is to get the best possible users from your User Acquisition efforts
Coordinate with UA efforts to find these users. Identify them based on:
• Source• Demographics• Look-a-like
targeting• Behavioral
characteristics
Level of Usage
Valu
e of
a U
ser
PRACTICAL IMPLEMENTATION
STAGES OF SEGMENTATION
• Leverage more data to build predictive modeling and advanced statistical analysis
• Smaller, more-specific segments help stretch users to drive behavior that is incremental
• User-level segmentation
• Segment based on app usage (opens, purchases, etc.)
• Bucket users together into a manageable number of segments
• All users receive a uniform experience
• Useful for A/B testing general app features
None AdvancedIntermediateBasic
Example:
None Usage + Value Scoring Buckets
User-Level Usage + Value Scoring
K-means cluster analysis / Machine
Learning
1ST STAGE OF SEGMENTATION
Segmentation parameters should be aligned with business objectives For example: If ads are what make you money, segmentation should drive more ad
revenue The goal is to move users up and to the right
No Segmentati
on
High ValueLow Use
High ValueHigh Use
Low ValueLow Use
Low ValueHigh Use
High ValueLow Use
High ValueMid Use
High ValueHigh Use
Mid ValueLow Use
Mid ValueMid Use
Mid ValueHigh Use
Low ValueLow Use
Low ValueMid Use
Low ValueHigh Use
Level of Usage
Valu
e of
a U
ser
Valu
e of
a U
ser
Level of Usage
2ND STAGE IN SEGMENTATION Use cluster analysis to identify user groups that behave similarly Cluster analysis leverages statistics to allow users to group according to
pre-determined behavioral factors (purchase events, app feature usage, etc.)
2ND STAGE IN SEGMENTATIONShrinking
Growing
New User
Steady Eddy
Paid UA ReferralOrganic UA
Download
Data Collection Source Time to Register Demographics
Name Age Gender Zip/State iOS/Android
Actionable Targeting Data
User Behavior(Opens, app events, etc.)
Cus
tom
er S
uppo
rt &
Res
olut
ionData Collection
Engagement Level & Timing(How long does it take a user to cycle through funnel)
Data Collection
Frequency & Volume(Avg 1.2 per week, 2 wk/month)
ADVANCED SEGMENTATION – CONTINUOUS CYCLE
Data Collection Time App Events Ad Engagement Clicks Customer
Support contact Details for each
engagement or behavior
Source Source
Needs: 1) User Linked Across all
Platforms2) Dynamic
Learning/Updates 3) Data Warehouse 4) Personalization Tools
Actionable Targeting Data
App Events
Ad Interaction
Content
Likely to Churn
Other Variables
Likely to be Power
Likely to Refer
Source
Internal External
Communications
In AppEmail Push Ads URUA
Data (Predictive, Manipulation, etc)
ADVANCED SEGMENTATION – CONTINUOUS CYCLE
Data, Content & Communication
Triangle
RISKS AND LIMITATIONS Risks:
Users notice differences and get frustrated Segmentation is too complex to manage efficiently Law of Unintended Consequences
Limitations: Don’t house or collect all of the right data Can’t align across internal and external tools Don’t have the capabilities in-house to build
RESULTS FROM INCREASED SEGMENTATION
None Basic+72%
Increase in App Events
Basic Intermediate
+44%
Increase in App Events
KEY TAKEAWAYS TEST EVERY CHANGE IN SEGMENTATION!
Improve segmentation incrementally (and know what every improvement is worth)
Align segmentation with business goals
Let the data dictate your strategy
Build back end systems with future segmentation in mind
MAKE SEGMENTATION A CORE COMPETENCY!