KPI
Report
Training
Communicate
Explain/Fix/What does it mean?
Justify you did a good jobY
our
Effo
rt
Videos(€)
Media Views
Library Media Views
Live Streams
Qty Reminde
r
If Event is in the future, would a‘Add to my Calendar’ button increase the number of people watching the Livestream?
How about a mobile Push Notification?
Add to My Calendar
Example – Media / Content
Example – BankingPurpose of Website
Increase € in Current/Savings Accounts
Current/Savings Account
s (€)
Qty Current/Savings Account
s
Qty Clicks
Compare
Table
Avg Amount
per Account
Products
Without Tables
Example – BankingPurpose of Website
Increase Share of Wallet of Customers with > €50K Savings
Increase # of Products per Customer of Customers with > €50K Savings
Share of Wallet
(€)
Qty Users
Investing
CR% TargettedEmail
Avg Amount Invested
Step 2 : Determine Performance Metrics
Step 1 : Why is the website there?
Step 4 : Build Performance Analytics Tree
Step 3 : Determine Initial KPIs per Performance Metric
Step 5 : Prioritize Performance Analytics Tree
Step 6 : Implement Performance Analytics Tree
Spencer AltmanHead of Business Consulting
Berlin, Germany
+49 (0)160 928 707 [email protected]
@spenceraltman
Amsterdam | Beijing | Berlin | Frankfurt | Madrid | Milanwebtrekk.com
Keys to Getting Value Out of Data
• Head Management in the Data
• Dedicated Analytics resource/team
• Clear & powerful KPI structure
• All teams in the value chain sit together
CEO
CFO COO
Category Managers
CIO CMO
Digital Marketing
Analytics Channels
Display
AdWords
Social
Affiliate
Retargetting
Trafficing
17 Total Channels
SEO Creative/ UX Mobile
Web Apps
iOS Phone
iOS Pad
Android
CRM
Data Mining
Customer Database
Customer Relations
Offline Marketing
Market Research
CEO
CFO COO
Category Managers
CIO CMO
Digital Marketing
Analytics Channels
Display
AdWords
Social
Affiliate
Retargetting
Trafficing
17 Total Channels
SEO Creative/ UX Mobile
Web Apps
iOS Phone
iOS Pad
Android
CRM
Data Mining
Customer Database
Customer Relations
Offline Marketing
Market Research
Color Key
Needs Reports/Analytics
No Need for Reports/Analytics
URM / CRM
Analytics
Creative / UX
Site
Man
agem
ent
Emai
l
SEO
Dis
play
AdW
ords
Soci
al
Affilia
te
Reta
rgetti
ng
Traffi
cing
All O
ther
Offs
ite C
hann
els
Mob
ile
Market Research
Offl
ine
Customer Relations
Potential Marketing Matrix Structure
Extended Marketing Matrix Structure
URM / CRM Category Managers
Analytics
Creative / UX
Market Research
Clot
hing
Acce
ssor
ies
Cate
gory
X
Customer Relations
All Channels (Online & Offline)
Analytics in the OrganizationRecommendation :
Each Marketing and Category team is trained on using Webtrekk and getting value from the Analytics data.
Plus, there should be at least one person per team who is designated as expert in Webtrekk system. This person is the main contact to the Reporting/Analytics team for the following questions impacting their team:
• Tagging• Training/Reporting• Analysis
CRM
Basic Training
+ 1 Expert
Site Mgmt
Basic Training
+ 1 Expert
Category X
Basic Training
+ 1 Expert
Team X
Basic Training
+ 1 Expert
Reporting /Analytics
Team
Analytics Team BreakdownResponsibility
Report Needs / Training of Other Groups / Report Generation
Tagging / Implementation Concept / Implementation Management
Analysis & Insights
Skill Need
Expert / Familiarity with Web Analytics and Digital Marketing Tools
(Recommended : 2 years experience)
Javascript/HTML/More Technical;Web Analytics tagging preferred
(Recommended : 2 years experience)
Business Analyst / Evangelist /Familiarity with Data Analysis more important
than Web Analytics Knowledge; Communicator with Top Managers
(Recommended : 5 years experience)
Recommendation : One person per each responsibility. At the beginning of most projects, all 3 responsibilities are covered by a single person. This team ensures uniformity of analytics processes across the different teams using analytics data.
Spencer AltmanHead of Business Consulting
Berlin, Germany
+49 (0)160 928 707 [email protected]
@spenceraltman
Amsterdam | Beijing | Berlin | Frankfurt | Madrid | Milanwebtrekk.com
What is URM? Why URM?
Using RFM / RFE results
What is RFM / RFE?
How to configure RFM / RFE
Example of URM in Action
Agenda
URM : Driving Profit RETARGETING
CHURN PREDICTION
PURCHASE
SMART LANDINGPAGES
RECOMMENDATION
EXIT
RETURNING BUYERS
Analytics/ URM
Website
Display
Mobile (incl.Push)/In-Store
Social Media
Store
Call Center
Post/Paper
Retargetting
Website & Apps
CRM
Social Media
Other
Display
RetargettingDWH & Other
Systems
Marketing Automation
SearchSearch
Channel Overview
URM enables…
More new customers
1
More profit:
• per customer
• per transaction
• per visit
2 3
Reduce time
between transactions
Last Transaction
Last Visit
Next Transaction
Next Visit
Limited Resources Information Budget Interactions with the User/Customer
Optimize Available Resources Example: Generic Newsletter
Costs Response rate Lose relevance with your customers Dislike from your customers Destroy your e-mail sender rating
Be Relevant and Profitable
Why RFM/RFE?
Various models, just one example (Customer Life Time Value) High price and frequently returning customers
1 purchase for €1000
10 purchases for €100
100 purchases for €10
1000 purchases for €1 Variable and Fixed Costs
Customer Segmentation
What is the RFM Model?
Proven Scoring model for defining customer value
Based on historical customer data
Goals:
Prediction from purchase likelihood and return rates
Segmentation based on customer value
Target customers/users with right message
RFM Model
Recency
When was the last purchase?
Frequency
How often was purchased?
Monetary
What was the total order value?
Recency Score
Frequency Score
Monetary Score
RFM Score
RFM Model
Recency
When was the last visit?
Frequency
How often are the visits?
Engagement
What was the total engagement?
Recency Score
Frequency Score
Engagement Score
RFM Score
RFE Model
3 Rating Groups per section
2 threshold value split scores into the 3 groups
Exact limits are company and case specific!
Customers
Rating Group 1
Threshold 1 Threshold 2
Rating Group 2 Rating Group 3
RFM Model Scores Explained
Threshold 1 Threshold 2
1
bad
Recency Last purchase more than 90 days ago 90
Last purchase made between 30 and 90 days ago
30 Last purchase occurred within the last 30 days
Frequency Quantiy orders less than 3 3 Quantity orders
between 3 and 10 10 More than 10 Orders
Monetary Order value less than €100 100 Order value between
€100 – €1000 1000 Order value is higher than €1000
Score: 2
middle
3
good
Example RFM Model
1Define Goal : Limiting Wasted Marketing Spend or Improve Revenue
for Key Target Groups
3 Determine how to structure the part of curve you are to focus on
4 Go into URM data (existing standard configuration) to determine thresholds
5 Target campaigns to relevant segments via marketing actions
6 Verify results and adjust (a) RFM/RFE? or (b) marketing actions?
RFM/RFE Setup Overview
2 Determine to focus on head of curve or tail of curve
Action based Grouping:
Churn Prevention Use Churn time frame as lower limit (RFM and RFE)
Break Even Profitability Use Break Even Cost (approx.) as lower limit Use desired Profit Margin as upper limit
Power Shoppers/Users
Configuring the Model
RFM Group
MonetaryWhat Order-Value was reached?
1bad1
bad3
good3
good
FrequencyHow often did they purchase?
1bad1
bad3
good3
good
RecencyWhen was the last purchase?
2mid2
mid3
good3
good1
bad
Mobile-Phone without contract (IPhone)
2mid
2mid
Bought last phone 2 years ago
Bought a new phone every 2 years
IPhone 4s – best price value configuration
122
Examples of RFM Groups
FrequencyHow often did they purchase?
2mid
MonetaryWhat Order-Value was reached?
2mid
2mid
1bad
1bad
3good3
good
RFM Group
3good3
good
RecencyWhen was the last purchase?
2mid2
mid3
good3
good1
bad
Mobile-Phone without contract (IPhone)
2mid
Bought last phone 2 years ago
Only one purchase
IPhone 4 – minimal-features
111
Examples of RFM Groups
FrequencyHow often did they purchase?
1bad1
bad
MonetaryWhat Order-Value was reached?
3good
3good
RecencyWhen was the last purchase?
1bad
2mid
2mid
2mid
1bad
RFM Group
2mid2
mid3
good1
bad
Mobile-Phone without contract (IPhone)
2mid
Purchase on release
every model will be purchased
IPhone 6s – top-features
333
1bad
Examples of RFM Groups
Actions: Loyalty Programm Thanks, Personalised, Spread
Low Frequency High
ThanksHi
gh Continue UpsellImprove
Cross-Sell
Recency Increase Guide
1 time
Low Do not
continueActivate
Reactivate with Upsell
Reactivate
MonetaryLow High
Possible recommended actions for RFM segments
How to approach customers?
Activate
RFM Group 123
target segment
Visitors New Visitors Bouncer Returning Visitors
Female Age Cluster:18-25 Years
Predefined / custom segments
How to approach customers?
address them
RFM Group 123
Better Targeting and Actions:
Expand Segmentation to address segments even better Predefined:
URM Customer Micro/Macro Status, Conversion Probability, Churn Probability, Next Basket Value, Age, Gender, CLV, Qty Product Views
Custom: Product Return Rates, Household Income
Include "offline contacts" to keep model accurate Leverage channel preference
Cases Create Communication channel for "new" users
Push the sale or create a connection? Post Sale Experience – Service Offering Targeted On Site Survey Preferred Marketing Channel Call Center Integration
Using RFM/RFE Plus X
Possible approach:
Churn Prevention Use Churn time frame as lower limit for Frequency Target people onsite before churn
Offer (Premium) Content to right audience Incentivize Log-ins Personalized Recommendations
No Products to sell? - Expand Model to Ad clicks Work with performance based marketing for higher eCPM
Increase Premium Ad Prices (combination with URM information)
Establish a persistent communication channel
Using RFM/RFE for Publishing
Begin to use URM / MA
Marketing Automation
UserRelationshipManagement
Because the user searches with an iPad, we show higher priced
items, such as :Leather Sofas
Campaign Channel: SEOKeyword: sofa
Device: iPad
Begin to use CRM Data
Marketing Automation
UserRelationshipManagement
Indications of customer who buys higher priced
items
Thus, even higher priced
recommendations than just leather sofas
Campaign Channel: SEOKeyword: sofaDevice: iPad
Gender: Woman
Income: €60K+
More CRM Data
Marketing Automation
UserRelationshipManagement
With children in the house,
maybe leather/too many high priced
items is not such a good idea
Campaign Channel: SEOKeyword: sofaDevice: iPadGender: WomanIncome: €60K+
Kids: 1
Include Survey Data
Marketing Automation
UserRelationshipManagement
It turns out the child is likely already in or
moving to university
If it (significantly) impacts the likelihood to buy, we should ask the
customer…
Campaign Channel: SEOKeyword: sofaDevice: iPadGender: WomanIncome: €60K+Kids: 1
Next Big Project:
Student Flat
Example : Potential Ikea Survey
For a 10% discount off of your next purchase over €1.000, please tell us your next big project:
New KitchenNew Living RoomNew BedroomApartment for our childSecond Home
Customer Looks at a Specific Product
Customer browses the site for various sofas
and sofa beds, including the one to the left.
Customer does not purchase and leaves the
site.
What could happen next?
Targetted Email
Campaign Channel: SEOKeyword: sofaDevice: iPadGender: WomanIncome: €60K+Kids: 1Next Big Project: Student Flat
RFM:333Preferred Marketing:
The customer :
Bought Recently (R = 3)Buys a lot Frequently (F = 3)Spends a lot of money with us (M = 3)
Customer also purchases via our newsletter sometimes.
Therefore, we will close following marketing channels for the next 15 days for this customer:
• Criteo Retargetting• Display• Affiliate• AdWords
Focus on Email for 15 days.
Instore Push Message
Customer looked at
sofaonline
One Week Ago Today in Store
This sofa is just a 2
minute walk away in the sofa
section
Integration Overview from Examples
Analytics/ URM
Website
Mobile (incl.Push)/In-Store
Criteo
Display
AdWords
Affiliate
Website & Apps
CRM
AdWords
Affiliate
Display
RetargettingDWH & Other
Systems
Marketing Automation
Survey
Conrad MorbitzerSenior Business Consultant
Berlin, Germany
+49 (0)30 755 415 [email protected]
Amsterdam | Beijing | Berlin | Frankfurt | Madrid | Milanwebtrekk.com
MA / URM Quick Start
UserRelationshipManagement
1
CreateUser Segments
Marketing Automation
2
LaunchCampaigns
3
AutomatedPersonalised
Banners & Content
Customer Prints Product
Form
Without Webtrekk User Centric – Data shows 2 People
Today
Customer Submits
Form/Signs Up for Product
In 6 Weeks
Customer Prints Product
Form
Webtrekk User Centric – Data Shows Single Customer
Today
Customer Submits
Form/Signs Up for Product
In 6 Weeks
Bank Example 1 – New Product Signup
Customer Prints Product
Form
More Targeting, Shorter Cycles
Today
Customer Submits
Form/Signs Up for Product
In 6 2 Weeks
Only to customers with Account
holdings greater than
€XXX.000
Bank Example 2 – New Product Signup
Customer Plays with Mortgage Calculator
Customer Signs up for
Home Loan in a Branch
Click here to set up an appointment to take
advantage of the best mortgage rates in
History.
Bank Example 3 – Home Loan
?Unknown prospect
searches for ‘best savings
account‘
Unknown
Visitor
Searches
Compare savings options
Transfer nowReceive 2%
Banners on SitePersonalisation
LocationBerlin
LocationMunich
DeviceiOS
DeviceAndroid
HighlyEngaged
Visitor
LowEngaged
Visitor
Bank Example 4 – Prospect Handling
Customer looks at dress online
Last Night Today
This dress is
available at the front
of the store near
shoes
E-commerce Example