9/20/2011
1
Measuring Customer Value: A Different Perspective on Price Optimization
October 3 – 4, 2011CAS I F S iCAS In Focus Seminar
Roosevelt C. Mosley, Jr., FCAS, MAAA
Pinnacle Actuarial Resources, Inc.
Experience the Pinnacle Difference!
Measuring Customer ValuePrice optimization Customer value
Interest and use is growingDefinition can be challenging
Narrows the lens to the exclusion of valuable inputs
Components (profitability, customer response)CompetitionElasticity of demand
Applications beyond pricing are limited ApplicationsPractical considerations
9/20/2011
2
What is Price Optimization?
Integration of demand side pricing into an Google Definitions
overall pricing strategyOffering each customer the right price to achieve pricing and profit goalsThe process of setting prices to maximize a
d fi d f lpre-defined measure of customer value subject to a company’s strategic and business objectives
Definition
OptimizeOptimize1. to make as effective, perfect, or useful as
possible2. to make the best ofOptimization: a mathematical technique for finding aOptimization: a mathematical technique for finding a maximum or minimum value of a function of several variables subject to a set of constraints, as linear programming or systems analysis
Source: www.dictionary.com
9/20/2011
3
Expanded DefinitionXXXX optimization: mathematical technique for finding the best XXXX (most effective, perfect, most useful maximum minimum) subject to a set ofuseful, maximum, minimum) subject to a set of constraintsXXXX =
PriceMarketing strategyCl i ttl tClaim settlement processClaim fraud identificationUnderwriting processPrice change
Expected Profit Curve
Optimal Profit Actually profit that is optimized
subject to price
Expected Prof
it subject to price and other
constraints.
Price
9/20/2011
4
Constraints
Profitability: corporate goalsGrowth
Success of marketing effortsNew business hit ratios
Loyalty: retention ratiosCompetitive: competitor price AND operational considerationsFuture: lifetime value implicationsA i l i i d f i lActuarial: not excessive, not inadequate, not unfairly discriminatoryRegulatory: must operate within parameters of the law
Limits of Price Optimization
Focus is necessarily on priceWhat am I really optimizing?What am I really optimizing?Practical implications of price optimization are unclearApplication to other areas of a company is a challengechallengeWhat happens when the optimal price is significantly different than the price I can charge?
9/20/2011
5
Expanding the Focus
C stomerPricingCross Sell
Marketing QuoteUnderwriting
SaleOngoing Servicing
Customer Service
RetentionClaims
Re-underwritingScenarios1. Price Change Only2. Price, New Business and Renewal Changes3. Complete System Changes
Defining Customer Value
Instead of focusing solely on price, expand the lens to including pricing asexpand the lens to including pricing as
well as broader company considerations
9/20/2011
6
What is a Valuable Customer?
Low claim frequency risksNiche customerCustomer segment where rates are adequateCustomer segment a company is / is not successful writing Measures: historical or short term, aggregate
Issues With Current Customer Value Definitions
Can be subjectiveShort term viewShort term view
Ignores longer term potential valueComplicates long term changes
Can hinder implementation of more advanced insurance techniques
Pricing advancesOptimization
9/20/2011
7
Customer Value Definition
P (1 C ) R P (1 C ) R P (1 C )Expected Value of Existing BusinessEVEB = P1 x (1 – C1) ------------------------(1 + r) R1 x P2 x (1 – C2) --------------------------------(1 + r)2+ R2 x P3 x (1 – C3) --------------------------------(1 + r)3Pi = profit at time i = Premium – E(Loss) - ExpenseCi = Probability of cancellation during period i
+
Ri = Probability of renewal at the end of period ir = discount rate
Potential Customer Value Definition
Value of Quoted Business: VQB = VEB x P(Co)Q Q ( )
VTB = P(Q) x VQB ---------------------------(1 + r)tlValue of Targeted Business
P(Co) = probability of converting quoted business
(1 + r)tlP(Q) = probability of targeted business receiving a quotetl = time lag between being targeted and receiving a quote
9/20/2011
8
Elements of Insurance Customer Value
Elements of Insurance Customer Value
Expected profitCustomer Response (Likelihood of Retention/Conversion/Marketing Success)Influence of CompetitionInfluence of Price Changes
9/20/2011
9
Expected Profit
Premium set by the company
E(Profit) = Premium – E(Cost)Premium - set by the companyE(Cost)
For setting premiumsModeled loss costs by techniques accepted by industry and regulatorsVariables used limited by regulatory/public acceptance, operational considerations
For purposes of determining E(Profit)For purposes of determining E(Profit)More advanced techniques can be usedList of potential variables is “unlimited”
Modeling Expected Cost – Expanded Considerations
Use All Available Information Use Additional/Advanced Techniques
RatingUnderwritingMarketingAgencyBillingExternal (demographic, etc.)Endorsement activity
Decision treesNeural networksClusteringPrincipal componentsAssociation analysis
Endorsement activityPayment historyPortfolio information
9/20/2011
10
Examples of Additional Significant Cost Variables –Auto Analysis
Presence of additional lines (home, commercial, health)Additional line loss experienceScheduled propertyAmount of insuranceAge of homePrior property claimsPrior property claimsHome occupancy typeUmbrella indicator
Auto Indicated Claim Frequency
9/20/2011
11
Comprehensive Severity Model Comparisons
Linearity creates issues at the extremes
Expenses
Some are difficult to vary by riskSubtract fixed dollar amount – can change relativeSubtract fixed dollar amount can change relative profitability but maintains orderSubtract fixed percentage of premium or expected loss – maintains relative probability
Expenses that could vary by riskG l l iGeneral claim expensesMarketing expensesCustomer service expenses
9/20/2011
12
Difference Between “Full Model” and “Rating Plan Model”
18.9% 15.3%
Distribution of Expected Profit
9/20/2011
13
Customer Response Process
Marketing Effort QuoteSaleRenewal QuoteRenewal
Customer Response Models
Quoting AnalysisA l i f h lik lih d f i i d b i i i fConversion AnalysisAnalysis of the likelihood of a insured that has received a quote purchasing insurance from you
Analysis of the likelihood of a prospective insured obtaining an insurance quote from you
Retention AnalysisAnalysis of the likelihood of a current insured renewing with you
9/20/2011
14
Quoting AnalysisAnalyze characteristics of shoppers and quotersCharacteristics
l i f iInternal company informationExternal demographic information
Credit profilesMarketing profiles
Identify insureds to targetCan be done at different levels (individual target, geographic area, risk segment target)
One Way Summary – Means of Entry
9/20/2011
15
Quoting Analysis Example – Vehicle Year
Response Percentage
9/20/2011
16
Retention/Conversion CharacteristicsTraditional Rating Factors
ClassTerritoryLimit
Agent/Distribution Channel Issues
Satisfaction with Agent/ServiceLimit
Insurance ScoreClaims historyViolation history
Account CharacteristicsNumber of Years InsuredNumber of PoliciesAccount Size
Agent/ServiceDistance to AgentIndependent vs. Captive vs. Direct
Market ConditionsCompetitive PositionInflationU/W CycleRenewal Pricing Change /
New Business Price Difference
U/W CycleBrand Value (company & competitors)
Conversion Analysis Example
Example
1.000 1.000
1.408 1.431 1.483
1.766
0.800
1.000
1.200
1.400
1.600
1.800
2.000
Rel
ativ
it
0.000
0.200
0.400
0.600
<missing> 0 1 2 4 6EDUCATION
Relative Likelihood
9/20/2011
17
Retention Analysis
Influences on Customer Value
Price changesImpacts expected profitp p pAlso impacts expected customer conversion and retentionInteraction of these elements can produces unanticipated changes in customer value
Competitive positionInfluences ability to write/retain risksUltimately impacts premiums chargedDepends on more than just priceChanges over time and varies by type of risk
9/20/2011
18
Insurance DemandInsurance Demand Curve
0.70
0.80
0.90
1.00
Retention by Age
0.00
0.10
0.20
0.30
0.40
0.50
0.60
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
Percent Change
P(R
enew
al)
Retention by Age
0.6
0.8
1
1.2
tion
Like
lihoo
d
0
0.2
0.4
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
Rate Change Band
Rete
nt
LT 26 26-49 50-64 65+
Competitive Influence on Price Impacts
SegmentedMarketShareLife CycleAnalysis
CompetitiveQuotesMarketPosition
•Based on market realitiesBatch QuoteAnalysis
•Still retains analysis of price•Also includes market reality of what is being written/retained•Allows insurer to then investigate cause of competitive issues
9/20/2011
19
Measure of Influence of Competition
Com petitive Index
0.88
0.88
0.77
0.66
4
5
6
7
1.03
1.02
0.90
1
2
3
Analysis of Batch Competitive Quotes
Due to competitive analysis difficulties, batch quoting is not an exact scienceAssumptions
Because of insurance scoring and other proprietary elements, absolute rate level is difficult to determineMore likely to get relative rates between risks correct
Use batch quotes to determine “effective” competitor relativities – what is the estimated rate pdifference between your rate and rate of other companies for risk characteristicsUse batch quotes to get a range of competitor prices
9/20/2011
20
Credit Scores Conversion
Example
0 801.001.201.401.601.802.00
Rel
ativ
ity
0.000.200.400.600.80
Insurance Score
R
Detailed Market Share Segments
1. Detailed vehicle/property market penetration by:Geography (ZIP, census block)V hi l h t i ti ( d l k / d l hi lVehicle characteristics (model year, make/model, vehicle segment)Home characteristics (year built, AOI, construction)Driver characteristics (age, marital status, gender, occupation)
2. Penetration levels based on demographic characteristics
3 Penetration levels by company risk characteristics3. Penetration levels by company risk characteristics4. Models of projected market penetration based on
historical penetration data
9/20/2011
21
Penetration by Model YearABC Insurance Company
3.50%
4.00%
80.0%
90.0%
100.0%
1.00%
1.50%
2.00%
2.50%
3.00%
Mar
ket P
enet
ratio
n
20 0%
30.0%
40.0%
50.0%
60.0%
70.0%
Incu
rred
Los
s R
atio
0.00%
0.50%
1973
& O
LDER
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Vehicle Year
0.0%
10.0%
20.0%
Market Penetration Incurred Loss Ratio
Putting It All Together
P (1 C ) R P (1 C ) R P (1 C )Value of Existing BusinessVEB = P1 x (1 – C1) ------------------------(1 + r) R1 x P2 x (1 – C2) --------------------------------(1 + r)2+ R2 x P3 x (1 – C3) --------------------------------(1 + r)3Pi = profit at time i = Premium – E(Loss) - ExpenseCi = Probability of cancellation during period i
+
Ri = Probability of renewal at the end of period ir = discount rate
9/20/2011
22
Modeling Expected Cost -Considerations
Modeling multiple yearsLife changes (household demographics)Life changes (household demographics)Aging of insuredsVehicle changesUncertainty and time value of money
Current Customer Value
9/20/2011
23
Customer Value at t = 3
Applications of Customer Value
9/20/2011
24
Customer Value Formulas
EVEB = P1 x (1 – C1) ------------------------ R1 x P2 x (1 – C2) --------------------------------+ R2 x P3 x (1 – C3) --------------------------------+Expected Value of Existing BusinessEVEB (1 + r) (1 + r)2 (1 + r)3+Value of Quoted Business: VQB = VEB x P(Co)Value of Targeted Business
VTB = P(Q) x VQB ---------------------------(1 + r)tl
Driving Customer ValuePremiums: actuarial
Cl i t l i d itiClaim cost: claims, underwritingCancelation rate: actuarial, customer service, agency, claimsRenewal rate: actuarial, customer service, agency, claimsConversion rate: actuarial, customer service, agency, ITResponse rate: marketing
9/20/2011
25
Application of Customer Value Measures
UnderwritingHistorically focused efforts on “bad” risks: high frequency, prior violations and accidentsUse customer value to Improve underwriting workflow, shift focus to less profitable risks
ActuarialPrice optimization focused on solely changing price to improve profitabilityCustomer value considers price as part of bigger picture to help drive customer value
MarketinggHistorical focus on hit ratio, how many customers are respondingUse customer value to develop target market profileHow do I better attract low response segments
Product management – how do all elements interact, common theme that flows through the entire program
Practical Considerations
Data/ModelingRegulatoryRegulatoryCorporate Culture
9/20/2011
26
Data/Modeling Considerations
Pricing ChangesAccount vs. Line/Subline ModelingAccount vs. Line/Subline ModelingNew Business – premium & quote informationCompetitive informationQuoting analysis
Regulatory Considerations
Pricing approachE(Loss) + Expense Load vs.( ) pPremium that optimizes profit
Market conductWhen you determine more profitable business, what do you do with it?
Underwriting?Marketing?Pricing?
Actuarial standards and codes
9/20/2011
27
Corporate Culture
Focus is generally on the short termBusiness value focuses on the longer termBusiness value focuses on the longer termMust carefully balance
Culture in many companies is that volume must be preserved at all costs
Business value introduces profitability at the individual insured levelFinds the optimal point that maximizes profit based on the combination of price and volume