The Mindset of Decision-Making:
Best Practices to Increase Agility
and Visibility with Operational
Decision Management (ODM)
Ryan Trollip
Practice Director, Decision Management
Agenda
 Decision Management & Automation
 Why now?
 Decisions and Process (BPM)
 Practical Decision Automation
 Innovation & Optimization
 Decision Automation Example
Decision Management
& Automation
3
Decisions Everywhere
Horizontal: best/appropriate price, cross-sell/ up-sell recommendations, loyalty promotions,
exception identification, risk/fraud assessment, straight-through processing approvals
 Claim
• Validation
• STP approval
• Exception routing
 Policy/Underwriting
• Eligibility
• Risk
• Pricing
 Annuity
• Recommendation
• Commissioning
• Payout calc.
 Loan
• Eligibility
• Risk
• Pricing
 Account
• Cross-sell
• Fraud/Alerts
 Credit Card
• Mkg Offers
• Fraud
• Credit limit
 Patient Care
• Drug interaction risk
warnings
• Follow-up alerts
 Member
• Services
recommendation
• Eligibility
• Benefit calculation
 Provider
• Patient eligibility for
services
 Benefits
• Eligibility
• Calculations
 Tax Payer
• Classification
• Audit flagging
 Citizen
• Program(s)
recommendation
 Land/Permits
• Conveyance
processing
• Contract compliance
 Service Mgmt
• Service prioritization
• SLA alerts
• Maintenance alerts
• Order configuration
Insurance Banking Healthcare Government Energy/Util./Telco
Decision Management -
Discipline Vs. Technology
 Decision Identification/Creation
 Decision Modeling
 Decision Optimization
 Decision Automation
5
Why do I care?
 Decisions are your IP
 Costs money to manage change
 Slow to change
 Often not visible
 Dispersed & duplicated, Inconsistent
 Lack of clear ownership
6
Difficult to manage!
public class Application {
private Customer customers[];
private Customer goldCustomers[];
...
public void checkOrder() {
for (int i = 0; i < numCustomers; i++) {
Customer aCustomer = customers[i];
if (aCustomer.checkIfGold()) {
numGoldCustomers++;
goldCustomers[numGoldCustomers] = aCustomer;
if (aCustomer.getCurrentOrder().getAmount() > 100000)
aCustomer.setSpecialDiscount (0.05);
}
}
}
Simple & Visible
Centralize
9
public class Application {
private Customer customers[];
private Customer
goldCustomers[];
...
public void checkOrder() {
for (int i = 0; i <
numCustomers; i++) {
Customer aCustomer =
customers[i];
if (aCustomer.checkIfGold())
{
numGoldCustomers++;
goldCustomers[numGoldCustomers]
= aCustomer;
if
(aCustomer.getCurrentOrder().ge
tAmount() > 100000)
aCustomer.setSpecialDiscount
(0.05);
}
}
}
Why now?
10
Nexus of forces
Mobile Financial App
Sign in to mobile financial app to access
personal financial data
In-context social recommendations
See social recommendations in financial
app for banks. Initiate account opening
from app
Secure cloud-based personal data
Financial app populates cloud based
personal financial data into bank’s
account opening steps
Customer propensity with big data insight
Bank combines customer data with big data benchmarks on
external data to determine customer propensity for specific
products offered during account opening – i.e. offer money
market, credit cards?
Rules for credit limits & up-sell
Automatic decision points on credit
limits and approval levels. Determine
post account opening offers – migrate
401K account transfers?
Process for approvals & orchestration
Streamlined process to combine activation
and approval routing for recommended
products – checking accounts, overdraft
protection line of credit and money market
Account is activated and funded
before customer logs off
Account opening with
The gap in business operations is widening
Time to open a new account in retail banking
5 minutes – best
76 minutes – worst
Availability of loan funds
2 days – best
37 days – worst
Time for conditional mortgage approval
15 minutes – best in class
4 hours - median
Source: Boston Consulting Group, The “New New Normal” in Retail Banking, 2012
Drivers to Externalize Rules
 Maturity of technology
 Nexus of forces
 Customer expectations
 Competition
 Customer Centricity
13
Decisions in Process
(BPM & Workflow Automation)
14
Simplify Business Processes
 Complexity leads to
 Us doing the wrong things
 Losing sight of the customer
 The customer losing interest in us
“Simplicity is the ultimate sophistication” Leonardo da Vinci
 Simplifying business processes means enabling and
managing business decisions
 Extract business rules from business processes
 Encapsulate business rules in decision models
 Enable ownership of processes and decisions
Accept low-risk
applicant
Decline high-
risk applicant
Process
medium-risk
applicant
Age 21 to 75
Age 16 to 21
Age over 75
Accept low-risk
applicant
Decline high-
risk applicant
Process
medium-risk
applicant
Previous Accident,
Existing Infraction
Good Record
Age 21 to 75
Age 16 to 21
Age over 75
Accept low-risk
applicant
Decline high-
risk applicant
Process
medium-risk
applicant
Previous Accident,
Existing Infraction
Good Record
New Customer
Long-standing Customer
1 or fewer Claims Multiple Claims
Age 21 to 75
Age 16 to 21
Age over 75
Accept low-risk
applicant
Decline high-
risk applicant
Process
medium-risk
applicant
Previous Accident,
Existing Infraction
Good Record
New Customer
Long-standing Customer
1 or fewer Claims Multiple Claims
Car Type
Custom Car
Sports Car
Age 21 to 75
Age 16 to 21
Age over 75
Accept low-risk
applicant
Decline high-
risk applicant
Process
medium-risk
applicant
Previous Accident,
Existing Infraction
Good Record
New Customer
Long-standing Customer
1 or fewer Claims Multiple Claims
Car Type
Custom Car
Sports Car
Decisions in process/apps cause
complexity
Age 21 to 75
Age 16 to 21
Age over 75
Process low-
risk applicant
Process high-
risk applicant
Process
medium-risk
applicant
Low risk
Medium risk
High risk
Determine
applicant type
Identifying decisions simplifies them
Process Simplicity
22
Practical Decision Automation
23
Signals for Decision Management
 Scale
 Agility
 Transparency
 Ownership
 Precision & Consistency
Which Decisions?
 Repeatable
 Non-Trivial
 Measureable Business Impact
 Frequency of change
Automation Complexity
If x = 1
Then Y = X
Manual
Decisions
Expert
Decisions
Simple
Decisions
Decision Words
 Determine if a customer is eligible for a product
 Validate the completeness of an application
 Calculate the discount for an order
 Assess which driver is high risk
 Select the terms for a loan
 Choose which loan to Fast Track
Innovate & Optimize
28
Use Analytics to Improve
30
31
Why? - Prescriptive
Deeper Analysis of Trends & Patterns
• Analysis and Exploration
• Ad-hoc Query
• Trend and Statistical Analysis
32
What should we be doing? Predictive
Foresight to Plan & Allocate Resources
• What-If Analysis
• Predictive Analysis
ANALYTIC-DRIVEN ORGANIZATIONS are distinguished
by their ability to leverage …
All perspectives
Past (historical, aggregated)
Present (real-time)
Future (predictive)
At the point
of impact
All decisions
Major and minor
Strategic and tactical
Routine and exceptions
Manual and automated
All information
All information
Transaction data
Application data
Machine data
Social data
Enterprise content
All people
All departments
Experts and non-experts
Executives and employees
Partners and customers
33
All Processes
Operational
Management
Etc.
34
Optimization
Decision Automation Example
35
Reference Architecture
 Centralize Decision & Workflow
 Cache for performance (if necessary)
 SOA backbone
36
Decision
Management
Workflow
/ BPM
ESB / SOA
MDM
Cache
System System System
Customer Behavior & Cross Sell
 Channels have different answers.. Sometimes conflicting.
 Inability to capture & act on
 interesting behavior.
 events, risk, etc.
37
Web
Call
Center
Branch
Channels
Customer Cross-Sell
Capture Predict Act
Customer Data
• Demographics
• Account Activity
• Product Holdings
• Channel Activity
• Information Requests
• Complaints
• …
Campaign Data
• Contact history
• Response/purchases
• Test campaigns
• …
Analyses
Predict who is likely to
respond, based on their
customer profile when
receiving the campaign
Scoring
Marketing
campaign
process
Key
Performance
Predictors and
Campaign
Results
Rank best 3
offers
Attitudinal Data
• Customer Surveys
• Twitter
• Discussion Forums
• Blogs
• …
Website
recommendation
engine
Sales campaigns
Use Case - Customer Behavior - Interest
 Customer hits information pages on the site
 Web interest captured in product/service B
39
Decision
Management
Workflow
ESB / SOA
MDM
Cache
Web
Call
Center
Branch
A
B
C
D
…
B
A
C
D
…
40
Use Case – External Events
 External events change internal priorities based on
rules/model
40
Decision
Management
Workflow
ESB / SOA
MDM
Cache
Baby
Web
Call
Center
Branch
A
B
C
D
…
B
C
A
D
…
41
Use Case – Call Center
 Not interested in certain products
 Or tell me about it later
41
Decision
Management
Workflow
ESB / SOA
MDM
Cache
Web
Call
Center
Branch
B
C
A
D
…
B
C
A
D
…
6 Months
42
Use Case – 6 Months Later
 Workflow (BPM) kicked off by event
 Customer called
 Closed deal, update cache
42
Decision
Management
Workflow
ESB / SOA
MDM
Cache
Web
Call
Center
Branch
43
Improve
 Which product is more attractive to a wealthy customer?
 Who do we risk loosing most as a customer?
43
Decision
Management
Workflow
ESB / SOA
MDM
Cache
Web
Call
Center
Branch
BIG
Data
Approach
 Identify
 Analytics and other techniques to identify what needs improving
 Model
 Models visualize, simplify and clarify business understanding
 Automate
 Automate complex decisions, processes and models
 Manage
 Automation is great… But the more complex the more difficult to
manage/change
 Specialization of systems workflow, Modeling, BRMS, CRM, ECM etc.
 Good management of decisions and process is key to flexibility and efficiency
 Improve
 Analytics used to improve processes and decisions
44
 5 things to remember about Decision Management:
 Decisions are independent of process
 Brings agility, visibility, consistency, centralization of decisions
 Makes process simpler & smarter
 Decouples decisions from the development lifecycle
 Allows the business to manage their own decisions
Wrap-up
Next Steps
 See Operational Decision & Process Management in action
 Learn what’s possible with Prolifics Offerings & Combination
offering By Prolifics & Decision Management Solutions
 Discovery Workshop
 Application Assessment
 Iteration 1
 Decision Modeling Training
 Implement a 10 week quick start
46
Ryan Trollip
practice director, decision management
prolifics | office: (646) 380-2895 | mobile: (774) 641-3666
rtrollip@prolifics.com | yahoo IM: rtrollip_prolifics

The Mindset of Decision-Making: Best Practices to Increase Agility and Visibility with Operational Decision Management (ODM)

  • 1.
    The Mindset ofDecision-Making: Best Practices to Increase Agility and Visibility with Operational Decision Management (ODM) Ryan Trollip Practice Director, Decision Management
  • 2.
    Agenda  Decision Management& Automation  Why now?  Decisions and Process (BPM)  Practical Decision Automation  Innovation & Optimization  Decision Automation Example
  • 3.
  • 4.
    Decisions Everywhere Horizontal: best/appropriateprice, cross-sell/ up-sell recommendations, loyalty promotions, exception identification, risk/fraud assessment, straight-through processing approvals  Claim • Validation • STP approval • Exception routing  Policy/Underwriting • Eligibility • Risk • Pricing  Annuity • Recommendation • Commissioning • Payout calc.  Loan • Eligibility • Risk • Pricing  Account • Cross-sell • Fraud/Alerts  Credit Card • Mkg Offers • Fraud • Credit limit  Patient Care • Drug interaction risk warnings • Follow-up alerts  Member • Services recommendation • Eligibility • Benefit calculation  Provider • Patient eligibility for services  Benefits • Eligibility • Calculations  Tax Payer • Classification • Audit flagging  Citizen • Program(s) recommendation  Land/Permits • Conveyance processing • Contract compliance  Service Mgmt • Service prioritization • SLA alerts • Maintenance alerts • Order configuration Insurance Banking Healthcare Government Energy/Util./Telco
  • 5.
    Decision Management - DisciplineVs. Technology  Decision Identification/Creation  Decision Modeling  Decision Optimization  Decision Automation 5
  • 6.
    Why do Icare?  Decisions are your IP  Costs money to manage change  Slow to change  Often not visible  Dispersed & duplicated, Inconsistent  Lack of clear ownership 6
  • 7.
    Difficult to manage! publicclass Application { private Customer customers[]; private Customer goldCustomers[]; ... public void checkOrder() { for (int i = 0; i < numCustomers; i++) { Customer aCustomer = customers[i]; if (aCustomer.checkIfGold()) { numGoldCustomers++; goldCustomers[numGoldCustomers] = aCustomer; if (aCustomer.getCurrentOrder().getAmount() > 100000) aCustomer.setSpecialDiscount (0.05); } } }
  • 8.
  • 9.
    Centralize 9 public class Application{ private Customer customers[]; private Customer goldCustomers[]; ... public void checkOrder() { for (int i = 0; i < numCustomers; i++) { Customer aCustomer = customers[i]; if (aCustomer.checkIfGold()) { numGoldCustomers++; goldCustomers[numGoldCustomers] = aCustomer; if (aCustomer.getCurrentOrder().ge tAmount() > 100000) aCustomer.setSpecialDiscount (0.05); } } }
  • 10.
  • 11.
    Nexus of forces MobileFinancial App Sign in to mobile financial app to access personal financial data In-context social recommendations See social recommendations in financial app for banks. Initiate account opening from app Secure cloud-based personal data Financial app populates cloud based personal financial data into bank’s account opening steps Customer propensity with big data insight Bank combines customer data with big data benchmarks on external data to determine customer propensity for specific products offered during account opening – i.e. offer money market, credit cards? Rules for credit limits & up-sell Automatic decision points on credit limits and approval levels. Determine post account opening offers – migrate 401K account transfers? Process for approvals & orchestration Streamlined process to combine activation and approval routing for recommended products – checking accounts, overdraft protection line of credit and money market Account is activated and funded before customer logs off Account opening with
  • 12.
    The gap inbusiness operations is widening Time to open a new account in retail banking 5 minutes – best 76 minutes – worst Availability of loan funds 2 days – best 37 days – worst Time for conditional mortgage approval 15 minutes – best in class 4 hours - median Source: Boston Consulting Group, The “New New Normal” in Retail Banking, 2012
  • 13.
    Drivers to ExternalizeRules  Maturity of technology  Nexus of forces  Customer expectations  Competition  Customer Centricity 13
  • 14.
    Decisions in Process (BPM& Workflow Automation) 14
  • 15.
    Simplify Business Processes Complexity leads to  Us doing the wrong things  Losing sight of the customer  The customer losing interest in us “Simplicity is the ultimate sophistication” Leonardo da Vinci  Simplifying business processes means enabling and managing business decisions  Extract business rules from business processes  Encapsulate business rules in decision models  Enable ownership of processes and decisions
  • 16.
    Accept low-risk applicant Decline high- riskapplicant Process medium-risk applicant Age 21 to 75 Age 16 to 21 Age over 75
  • 17.
    Accept low-risk applicant Decline high- riskapplicant Process medium-risk applicant Previous Accident, Existing Infraction Good Record Age 21 to 75 Age 16 to 21 Age over 75
  • 18.
    Accept low-risk applicant Decline high- riskapplicant Process medium-risk applicant Previous Accident, Existing Infraction Good Record New Customer Long-standing Customer 1 or fewer Claims Multiple Claims Age 21 to 75 Age 16 to 21 Age over 75
  • 19.
    Accept low-risk applicant Decline high- riskapplicant Process medium-risk applicant Previous Accident, Existing Infraction Good Record New Customer Long-standing Customer 1 or fewer Claims Multiple Claims Car Type Custom Car Sports Car Age 21 to 75 Age 16 to 21 Age over 75
  • 20.
    Accept low-risk applicant Decline high- riskapplicant Process medium-risk applicant Previous Accident, Existing Infraction Good Record New Customer Long-standing Customer 1 or fewer Claims Multiple Claims Car Type Custom Car Sports Car Decisions in process/apps cause complexity Age 21 to 75 Age 16 to 21 Age over 75
  • 21.
    Process low- risk applicant Processhigh- risk applicant Process medium-risk applicant Low risk Medium risk High risk Determine applicant type Identifying decisions simplifies them
  • 22.
  • 23.
  • 24.
    Signals for DecisionManagement  Scale  Agility  Transparency  Ownership  Precision & Consistency
  • 25.
    Which Decisions?  Repeatable Non-Trivial  Measureable Business Impact  Frequency of change
  • 26.
    Automation Complexity If x= 1 Then Y = X Manual Decisions Expert Decisions Simple Decisions
  • 27.
    Decision Words  Determineif a customer is eligible for a product  Validate the completeness of an application  Calculate the discount for an order  Assess which driver is high risk  Select the terms for a loan  Choose which loan to Fast Track
  • 28.
  • 30.
    Use Analytics toImprove 30
  • 31.
    31 Why? - Prescriptive DeeperAnalysis of Trends & Patterns • Analysis and Exploration • Ad-hoc Query • Trend and Statistical Analysis
  • 32.
    32 What should webe doing? Predictive Foresight to Plan & Allocate Resources • What-If Analysis • Predictive Analysis
  • 33.
    ANALYTIC-DRIVEN ORGANIZATIONS aredistinguished by their ability to leverage … All perspectives Past (historical, aggregated) Present (real-time) Future (predictive) At the point of impact All decisions Major and minor Strategic and tactical Routine and exceptions Manual and automated All information All information Transaction data Application data Machine data Social data Enterprise content All people All departments Experts and non-experts Executives and employees Partners and customers 33 All Processes Operational Management Etc.
  • 34.
  • 35.
  • 36.
    Reference Architecture  CentralizeDecision & Workflow  Cache for performance (if necessary)  SOA backbone 36 Decision Management Workflow / BPM ESB / SOA MDM Cache System System System
  • 37.
    Customer Behavior &Cross Sell  Channels have different answers.. Sometimes conflicting.  Inability to capture & act on  interesting behavior.  events, risk, etc. 37 Web Call Center Branch Channels
  • 38.
    Customer Cross-Sell Capture PredictAct Customer Data • Demographics • Account Activity • Product Holdings • Channel Activity • Information Requests • Complaints • … Campaign Data • Contact history • Response/purchases • Test campaigns • … Analyses Predict who is likely to respond, based on their customer profile when receiving the campaign Scoring Marketing campaign process Key Performance Predictors and Campaign Results Rank best 3 offers Attitudinal Data • Customer Surveys • Twitter • Discussion Forums • Blogs • … Website recommendation engine Sales campaigns
  • 39.
    Use Case -Customer Behavior - Interest  Customer hits information pages on the site  Web interest captured in product/service B 39 Decision Management Workflow ESB / SOA MDM Cache Web Call Center Branch A B C D … B A C D …
  • 40.
    40 Use Case –External Events  External events change internal priorities based on rules/model 40 Decision Management Workflow ESB / SOA MDM Cache Baby Web Call Center Branch A B C D … B C A D …
  • 41.
    41 Use Case –Call Center  Not interested in certain products  Or tell me about it later 41 Decision Management Workflow ESB / SOA MDM Cache Web Call Center Branch B C A D … B C A D … 6 Months
  • 42.
    42 Use Case –6 Months Later  Workflow (BPM) kicked off by event  Customer called  Closed deal, update cache 42 Decision Management Workflow ESB / SOA MDM Cache Web Call Center Branch
  • 43.
    43 Improve  Which productis more attractive to a wealthy customer?  Who do we risk loosing most as a customer? 43 Decision Management Workflow ESB / SOA MDM Cache Web Call Center Branch BIG Data
  • 44.
    Approach  Identify  Analyticsand other techniques to identify what needs improving  Model  Models visualize, simplify and clarify business understanding  Automate  Automate complex decisions, processes and models  Manage  Automation is great… But the more complex the more difficult to manage/change  Specialization of systems workflow, Modeling, BRMS, CRM, ECM etc.  Good management of decisions and process is key to flexibility and efficiency  Improve  Analytics used to improve processes and decisions 44
  • 45.
     5 thingsto remember about Decision Management:  Decisions are independent of process  Brings agility, visibility, consistency, centralization of decisions  Makes process simpler & smarter  Decouples decisions from the development lifecycle  Allows the business to manage their own decisions Wrap-up
  • 46.
    Next Steps  SeeOperational Decision & Process Management in action  Learn what’s possible with Prolifics Offerings & Combination offering By Prolifics & Decision Management Solutions  Discovery Workshop  Application Assessment  Iteration 1  Decision Modeling Training  Implement a 10 week quick start 46
  • 47.
    Ryan Trollip practice director,decision management prolifics | office: (646) 380-2895 | mobile: (774) 641-3666 rtrollip@prolifics.com | yahoo IM: rtrollip_prolifics