10. 10
USE-CASE:
Identify the root-causes why customers did not order after registration (specific countries)
Registration
started
Address
information
Order
created
REGISTRATION
ADDRESS
SELECTION
SHOPPING
REGISTRATION PROCESS &
ONSITE TRACKING
Extract
data and
push it to
Celonis
ANY
REGISTRATION
START
REGISTRATION
FINISH
ANY
DROP OFF
ORDER
CREATED
DATA VISUALIZATION
~70% ~30%
CHECKOUT / REGISTRATION PROCESS
11. 11
25%
22%
12%
8%
Payment Selection Pageview
Cart Pageview
Confirm Order
New Account Pageview
8%Address-Default overview pageview
4%Address - Edit/add home pageview
Exit Points
Next Steps Deep Dive Analysis
A-B Testing
CHECKOUT / REGISTRATION PROCESS
13. 13
Return in cep
network
Return order
created
Return refund
successful
Return in cep
network
Payment
reminder sent
900
50
Return order
created
850
Refunds
completed 850
Customer
return
Return in
Warehouse
Refund
RETURN PROCESS & BUSINESS
EVENT TRACKING
USE-CASE:
Identify optimization potentials in the return process (one specific shipping country).
DATA VISUALIZATION
Extract
data and
push it to
Celonis
RETURN PROCESS
14. 14
IMPACT
● Customer Contact
Cost
● NPS
● Potential financial
loss
ROOT CAUSE
Hypothesis:
● Parcel gets lost on
the way or in our
facilities
● Return order
created event is
not written
SOLUTION
Parcel delivered
to customer
Customer
returned: Return
in cep network
Return in
Warehouse
Manual refund
● Follow up project
with our Business
Excellence team
RETURN PROCESS - FINDING
16. 16
BUY INSCOPEDATA
● Knowledge about data
missing
● Data quality insufficient
● Too much data
● Long data processing
time
● Deep domain
knowledge missing
● Unclear expectations
● KPI to be influenced
missing
● Prioritisation conflicts
● Commitment from
project members
missing
● Management buy in
missing
CHALLENGES FOR ZALANDO
17. 17
SOLUTION APPROACHES
Zalando AWS account
Celonis
Application
Server
DB*
Data
Models
Pre-
Discover
Data
DATA DISCOVERY
*Mega Eventlog with billions of
datasets, different identifiers, …
PROJECT SETUP
● Moved from process discovery
towards very concrete questions
● Upfront defined KPIs which we
want to influence
● Put the stakeholder into the
drivers seat
19. 19
● ...use real data from the process itself
● ...can be combined with six sigma &
lean principles
● ...especially for process heavy topics
● ...fast where standard data connectors
to mining tool are available
POWERFUL TECHNIQUE
● ...right data (must fit your needs)
● ...data quality (complete and correct)
● ...tooling (mining & data processing)
● ...expertise (domain, data & mining)
PREREQUISITES