6. 6
What is personalization?
Personalization is the process of delivering relevant content to each segment of visitors.
A current member
would see:
While a prospect
would see:
“Click Here to Login to Your Account”
“Login to Dispute a Claim”
“Thank You For Your Loyalty! You are Eligible for Discounted Pricing”
“Never Worry about Your Bill Again - Enroll in Auto Pay”
“Register to Pay Your Bills Online”
“Register to Receive Monthly Promotions”
“We Offer the Most Competitive Pricing in the Market”
“Get 0% APR on Transfers and Purchases”
“Great Coverage at an Affordable Price!”
7. Nearly three-fourths (74%) of online consumers get
frustrated with websites when content (e.g. offers, ads,
promotions) appears that has nothing to do with their
interests. (Source: Janrain & Harris Interactive)
8. In a study of 650 multi-channel marketing campaigns,
personalized campaigns consistently and overwhelmingly
beat out static campaigns in generating a high response rate
from recipients. (Source: MindFire)
9. 9
Visitor Attributes used for Personalization
Unfinished Quote
Behavior
Click Behavior
First Time vs.
Repeat visitor
Context
Keyword
Search
engine
Referring URL
Device
Type
Time
Date
Referrer Type
Demography
Median Income
Median Age
Presence of Children
Locality Type
Education
Geography
City
DMA
Zip Code
State
Customer
data
Current vs. Prospect
Product
Type
Customer Loyalty status
Internal segment
name
• Searched for “Cheap Travel
Tickets to NYC” on Google
from an iPhone
• Likely Bachelors degree,
earning $75,000 - $100,000
• New customer
• Milwaukee
• First Time visitor
11. 11
var
Personalization Workflow
IDENTIFY DATA
FIELDS
CHOOSE DATA
FIELDS
DETERMINE
SEGMENTS
KPI FOCUS PER
SEGMENT
GENERATE IDEAS
Identify any and all
data/visitor attributes that
can be used to personalize
the experience (e.g.
device, referring source).
Focus on a single data field
at any given time in the
early stages of
personalization to help
keep things streamlined
and less complex.
Determine whether or not
there are any segments
associated with the data
field.
Determine if each segment
requires a separate goal or
not. Some segments may
share goals while others
may differ.
Generate test ideas for
each segment.
Paid – Non Brand
Paid – Brand
Social
Display
Social
Facebook
Instagram
Form Submits
Call to Quote
Align respective social
branding to landing page
Are visitors from Instagram
more willing to call?
TEST & PERSONALIZE
AB & Multivariate Test
Ideas
Test respective ideas for
each segment.
12. 12
Identify Data Fields
Trip Type
Trip Length
Identify any and all data/visitor
attributes that can be used to
personalize the experience (e.g.
device, referring source).
Referring Source
Length of Trip
Trip Type
Referring Source
13. 13
Identify Data Fields
CampaignID
Identify any and all data/visitor
attributes that can be used to
personalize the experience (e.g.
device, referring source).
Referring Source
CampaignID
New vs. Repeat
Referring Source
New vs. Repeat
14. 14
Choose an Initial Data Field: Referring Source
Focus on a single data field at
any given time in the early
stages of personalization to
help keep things streamlined
and less complex.
Leveraging analytics data will
help you identify the data field
that could benefit most.
236,445
Visitors
55%
Step 1 Rate
7%
Order Complete Rate
Paid Search
184,914
Visitors
49%
Step 1 Rate
11%
Order Complete Rate
Organic Search
38,021
Visitors
43%
Step 1 Rate
3%
Order Complete Rate
Display
12,730
Visitors
61%
Step 1 Rate
6%
Order Complete Rate
Email
16. 16
Generate Ideas
Non-branded search term visitors
May not be familiar with the company
More likely attracted to affordability
Testing Variables
Competitive messaging in the headline (ex: “One stop shop”)
Emphasize low cost in sub-headline
Targeted image
17. 17
Generate Ideas
Branded search term visitors
Familiar with the company
Know what they want
Testing Variables
Prominent login CTA
Remove informative content about the company from landing page &
emphasize CTA
18. 18
Testing & Personalization (Branded)
The test explored the
impact of streamlining the
landing page for branded
search visitors.
We saw an increase in
downstream conversions as
the challenger simplified
experience and focuses
more attention on primary
call to action.
0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
70.00%
Step 1 Order Complete
Control Challenger
+48.20%
Segment: Brand search term visitors
KPI: Step1, Order Complete
19. 19
Testing & Personalization (Non-Branded)
Segment: Non-brand search term visitors
KPI: Step1, Order Complete
7.30%
7.40%
7.50%
7.60%
7.70%
7.80%
7.90%
8.00%
8.10%
8.20%
8.30%
Order Complete UAR
Control Winner
+7.30%
This test explored the impact of
introducing informative
content to landing page on
non-branded visitor
conversion.
Additional product information
resonated with visitors
presumably less familiar with
the company & translated into
a positive lift in conversions
over the default experience.
21. Travel
• Destination/Arrival
• Day of week (weekday vs.
weekend traveler)
• Device (mobile vs. desktop)
• Discount code applied vs. not
applied
• Search results pages
• Add-ons page
• Geo specific Banners
• Order of search results
• Product/Search result badging
• Value driven headlines
• Recommended add-ons
• Reservation rate
• Average order value
• Add-on take-rate (think % of
people who take add-ons)
• Add-on Revenue
Common Data Fields Common Page Variables
Common Testing Iterations Common Metrics
22. 22
Travel Case Study
Segment: Weekday vs Weekend
KPI: Average Order Value, Reservation Rate
Goal/Hypothesis: Weekday renters, compared to weekend
renters, generally have a higher reservation rate. Test the
impact of adding badges to specific higher valued vehicles in
an attempt to increase average order value.
Test Variables: Blue badge above specific higher value
vehicles
• Take your trip from boring to fun
• Make this the best part of your trip and treat yourself
• Impress your clients with this car
• Upgrade your trip and travel in style
• Travel larger. This vehicle fits [insert #] passengers
+9%
Average Order Value
23. Common Data Fields Common Page Variables
Common Testing Iterations Common Metrics
• Device Type, location
• New vs. Return Visitors
• Current Customer vs New
• Last Product Ordered/
Browsing Patterns
• Search Box
• Homepage Product
Overview
• Shopping Cart
• Registration Page
• Promotions
• Seasonal Specific Product
• Security Messaging
• Membership Benefits
• Revenue per Visitor
• Conversion to Purchase
E-commerce
24. 24
E-commerce Case Study
Segment: Tablet
KPI: Revenue per Visitor, Thank You
Goal/Hypothesis: AB test to explore impact of new Tablet
design against Desktop design for Tablet visitors.
New experience personalized for a smaller screen size
resonated with visitors providing lift for both Revenue and
Thank You page conversion
Test Variables: Desktop vs new Tablet display for Tablet
visitors.
+285%
Revenue per Visitor
25. Insurance
Common Data Fields
• Device Type
• Location
• Demographic
• Insurance Type
Common Page Variables
• Homepage
• Product Content
Common Testing Iterations
• Insurance Quote CTA
• Informative Content
• Banner Demographic
Alignment
Common Metrics
• Locate Agent
• Call Agent
• Online Quote CTA
Engagement
• Quote Complete
26. 26
Insurance Case Study
Segment: Geo-location
KPI: Quote Start Clicks/ Contact Agent Clicks
Goal/Hypothesis: The goal of experiment was to explore
impact of introducing geo-targeted dual CTA (Quote/Agent)
within informative content pages.
Banner called out conversion methods available within the
visitor’s areas.
Test Variables: New mid-page banner with geo-targeted
options for both quote and agent.
+163%
Quote Starts
29. 29
Use available data points to identify and prioritize customer
segments to personalize against.
Analysis conducted of several different data sources
including
Analytics
Clickstream
voice of the customer
competitive data.
D
DATA
0.00%
5.00%
10.00%
15.00%
20.00%
25.00%
30.00%
35.00%
0
500
1000
1500
2000
Direct/Bookmarked External - Organic
search
External - Other Internal - Other
Referrer Type
Visitors Quote Start Thank You
30. 30
Consider allocation of resources as it pertains to every step
of the personalization process.
At the end of the day, you must be able to answer the
question - “Do I have enough resources to accomplish my
personalization goals?”
R
RESOURCES
Agency
Testing Vendor
IT Department
Personalize
• Develops
mocks
• Hosts assets
• Develops web
pages
• Analyzes data
points
• Generates ideas
• collaborates on
strategy
• Launches tests
• Manages pipeline
• Captures insights
• Declares winner
• Deploys segment
winners
• Manages release
schedule
31. 31
Identify areas of the page or website that could benefit
from testing and optimization and outlines
recommendations for how to improve performance.
Use your knowledge of the visitor segment to inform your
testing concepts. One size does not fit all as different
visitors behave differently and have varying reasons for
visiting your pages.
I
IDEATION
32. 32
Validate your personalized testing concepts through
AB/MVT testing
Even if your challenger/s fail to yield a positive impact, data
collected during experiment assists with identifying new
website problem areas and visitor drop off points.
V
VALIDATION
A/B
Vs.
MVT
33. 33
Establish a continuous testing process that allows your team
to prioritize pages and KPI through out the year
We shouldn't forget that the end goal is to deploy our
winning variations to our default content.
At the end of the day, testing should be iterative and our
insights cumulative. We should continuously build upon
previous results and learningsE
ESTABLISH