Post on 10-Aug-2015
Elmar Weber | @weberelm
Operational Process Analytics
Why traditional analytics and monitoringare not enough for todays BPM demands
Activiti User Day 2015, 10th June, Paris
About Me | Why Listen to My Crap?
• 8 Years consulting on BPM and big data projects in Java and Scala
• Worked for enterprises like SAP and EWE to solve
their BPM analytics problems
• Developed on a lot BPM solutions like SAP NetWeaver, IBM Process Server,
Oracle Fusion, Inubit, jBPM, Activiti,, …
• <shameless>
<plug>
Two years building software that specifically solves
BPM analytics problems
</plug>
</shameless>
Elmar Weber | @weberelm
Key Takeaways | If you don’t follow the whole 30 minutes
• Why operational intelligence and prediction matters
• Difference to technical monitoring
• Why pure BI doesn’t cut it
• How to get started with operational analytics on Activiti
No operational overview
back-office
High impactincident
Overdue
Taskbottleneck
Idle taskfor too long
Customercomplain
Manufacturingcontrol room
Shared Servicescontrol room
Workforce Management?
Incident Management?
Process Optimization?
Technical Monitoring vs. Business Monitoring
Business Problems
• Customer not reachable
• Supplier ran out of stock
• SLA about to get violated
Technical Problems
• Process crashed because
of infrastructure error
• Invalid data for gateway
expression evaluation
Business Questions
• Which ticket to solve first
• What action to take to prevent loss
• Impact of task times on costs
Technical Questions
• When to add new database server
• Technical task response times
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Impact towards Operational Excellence
Service Processes: History of BPM
MANAGEMENT
SCIENCE
WORKFLOW
AUTOMATION
BPM Principles
Frederic Taylor
Production Line
Henry Ford
Petri Nets
Carl Adam PetriTuring Machine
Alan Turing
INDUSTRIALIZATION
ANALYTICS
Internet of Things
Industrialization
4.0
BPM-Engines
Activiti, SAP, IBM, …
iBPMS
Appian, Pega, …
Business Intelligence
Business Objects, …
Operational Intelligence
SAP OPINT, …
SIX SIGMA
MotorolaLean Manufacturing
Toyota
Office Information Systems
Michael Zisman
Deep Learning
PredictionPalantir
Electrification
Industrialization
2.0
Digitization
Industrialization
3.0
Enterprise
Performance Management
IT Operational AnalyticsSplunk, New Relic, Sumo Logic
Simple-BPM
Effectiv, …
Business Intelligence vs. Operational Intelligence: So what’s the difference?
CRM PLMSRM HR SCM
Approve this
credit card
transaction?
Send this package
by air or ground? Expedite the order? What cross-sell
offer to make?
LEGACY
Adjust call center
operations?
How to recover
lost orders?
Is it time to order
more material?
Which customer
to target?
What policies for
credit card approval?
What prices to use
on our price list?
What product
specifications to use?
What business
should we be in?
Should we acquire
that company?
OPERATIONAL INTELLIGENCE
REGULAR BI AND ANALYTICS
Operational Decisions
Tactical Decisions
Strategic Decisions
Source: Gartner
Why traditional Business Intelligence can’t solve the problem (1)
Unstructured Data
Data Warehouse
SERVICE
LEGACYCRM
PLMSRM
HR
SCM
1 Connection
&
Synchronization
2Real-Time
3
Operational Context4
Pro-Active
&
Prediction
5
Call-to-Action
SHARED PROCESSES
SSC-AGENTS
CHALLENGE: RIGHT DATA, RIGHT CONTEXT AT THE RIGHT TIME!
Why traditional Business Intelligence can’t solve the problem (2)B
usi
ne
ss
Inte
llig
en
ce
LEGACY
CRM
PLM
SRM
HR
SCM
Data cubes
Queries
Forecasting
Real-time
Event based
Predictive
Reactive
Descriptive
Analytical
Proactive
Process context
Insight-to-action
Op
era
tio
na
l
Inte
llig
en
ce
Disrupt BPM through Operational Intelligence and Predictive Analytics
Contextual Data
Real-time monitoring (BAM)
Relevant Data
Predictive knowledge
OPERATIONAL ANALYTICS
PRO-ACTIVE DECISION SUPPORTC
U
P
E
N
Y
A
I
N
S
I
G
H
T
S
SSC-AGENTS
2
Business Events
+
Context Data
100% Data
in Real-Time
<->
1 Connector
3
1
KPI-Design
on-the-fly
LEGACYCRMPLM
SRM
HR
SCMERP-Layer
SSC
KPI-Templates
54
Proactive
Action-ListPrediction
+
Alerts+
+
SERVICESHARED PROCESSES
Easy Integration into Activiti
• Requirements:
Maven JAR Dependency and free cloud account
(com.cupenya.agent.activiti.cupenya-activiti-agent)
• Integrate BPMN Parse Listener into configuration (Spring, Standalone, …)
com.cupenya.agent.activiti.listener.CupenyaParseHandler
• Check it out:
https://github.com/cupenya/cupenya-activiti-agent
Outlook
• Process import
automatically discover and import deployed processes
• Insights to action
adjust process flow and outcome in Activiti
• Task State Support
add support for idle time, pending, etc based on Activiti task model
• Simulation (Crystallball)?
• ….
Predict, detect, act.www.cupenya.com
“Cupenya Insights truly does turn real-time business process monitoring into actionable insight, predicting operational issues and potential risks before they could manifest. And it’s really meant to be used by business people, because it was built as such – unlike the typical IT products with a pretty interface.
If you’re a manager looking to learn more about the operations of your business and find ways to optimize them, Cupenya Insights gives you everything you need. Also, it does that without the IT bottleneck and heavy investment associated with enterprise analytics.
And if you’re a large ‘frankenstack’ software vendor, then better watch your back.”
- THEO PRIESTLEY
Elmar Weber | @weberelm
elmar@cupenya.com
REAL TIME UPDATES
VISUAL KPI PERFORMANCE
MULTI-SYSTEM SUPPORT
DETAILED INCIDENT INFORMATION
CUPENYA INSIGHTS: NEXT GENERATION OF REAL-TIME PROCESS MONITORING
PROJECTED WORKFLOW PATH
CONTINUOUS MACHINE LEARNING
CUSTOMIZABLE NOTIFICATIONS
PREDICTED KPI PERFORMANCE
CUPENYA INSIGHTS: TICKET PREDICTION OF SLAs & PROCESS-ROUTING
CUSTOMIZABLE DASHBOARD
MANAGED PERIODIC DELIVERY
DRILL-DOWN REPORTS
CUPENYA INSIGHTS: SELF MANAGED ANALYTICS FOR REAL-TIME DASHBOARDS