june 16-17 • boston, ma turning data into a strategic asset · x44 tmc_amt x47 chronic x48 tds...
TRANSCRIPT
JPK
Gro
upOrganizational Intelligence Forum
June 16-17 • Boston, MA
Turning Data Into a Strategic AssetEmploy data analysis and metrics across every function inthe organization by creating a data/insight-driven culture
June 17, 9:45am
View presentation online at:
https://jpkgroupsummits.com/attendee4
Don Gray – Cigna
KANRI DISTANCE
CALCULATOR™
TURNING DATA INTO AN ASSET
Mov ing Beyond the Bas ics
June 17, 2016
OVERVIEW 1 The Basics
2 Beyond the Basics
3 Purpose-Driven Insights
4 Individualization
5 Actionable
THE BASICS
What the Business Gets
CURRENT “DATA AS AN ASSET” PARADIGM
KANRI Purpose-Driven Insights 4
Creating governed, authoritative data and ultimately master data
• Improved data trustworthiness and efficiency
• Increased persistency
Distributed metrics and performance measures under the banner of
business intelligence
• Run rate trend lines
• Period-over-period performance measures
Predictive Analytics and Population Data Management
• Fitting a population to a line, forecasting against formulas
• Identifying overall data drivers
• Population proof points
WHAT BUSINESS
LEADERS NEED
PURPOSE-DRIVEN INSIGHTS
DATA AS AN ASSET: Beyond the Basics 6
Analytics individual
level
Contributions of
variables affecting
individuals
Individualized
outputs help to
take decisions
that are affordable
1 INDIVIDUALIZATION
2 SPEED
Easy to interpret
analytical results
1 to 2 models per
week (much faster
than traditional
analytics)
Ability to add easily
new rows and
columns to prior
analyses
3 4
Distance
measures
Overall population
drivers of the
distance
Individual distance
drivers (variables
and combination of
variables)
EASE OF USE
Simple and easy
execution
Straightforward
outputs
Root cause
analysis
prioritization
through contextual
flagging
INSIGHTS PRECISION
1 to 2 models per
week (much faster
as compared
traditional
analytics)
Actionable insights
to partners
Enhanced quality
of insights
PRODUCTIVITY GAINS
5
Key Enablers: Simple & Affordable Innovative Technology, Tools Techniques and User Talent
Embrace and adapt to change faster
Deliver differentiated value through personalized results and distance metric
INDIVIDUALIZED
INSIGHTS
KANRI Purpose-Driven Insights 9
1. Define the target (healthy) group and all
attributes for consideration
2. Development of multivariate measurement
scale with target group as the reference
3. Calculate each off-target participant’s
distance from target by running data
through KDC™
4. Perform optimization as needed to reduce
the number of variables.
5. Perform individual diagnostics through root
cause analysis and identify contributing
variables for each participant
Importance of Correlations
KANRI Methodo logy
X1=Weight
X2=Height Target Group
KANRI
Distance
Sam
X
Raj
KANRI
Distance
Sam and Raj are close to target if we look
variables independently.
With correlations they are off-target. In fact
Sam has higher degree of abnormality
A N A LY S I S O F D I A B E T E S C A R E G A P S
Total # of variables – 83
# of variables considered for analysis - 49
Sample size
- Healthy – 11739
- Abnormal – 45227
- For prediction – 21564
• Optimized set of 32 variables
• Distances of each of the 45227 abnormals
and 21564 requiring prediction
• Contribution of each variable to abnormal
distances.
Data set Approach
Var Description
X2 age
X3 gender_cd
X5 New_Race
X7 prospective_risk_num
X13 TDS_lbp_flg
X14 TDS_cad_flg
X16 TDS_osteoknee_flg
X17 TDS_osteohip_flg
X19 TDS_bngnutrn_flg
X20 Wellness_prvntn_flg
X22 Wellness_hlthyeat_flg
X24 Wellness_hyprlpdm_flg
X25 Lifestyle_Weight_flg
X26 Lifestyle_Tobacco_flg
X27 Lifestyle_Stress_flg
X28 Chronic_Weight_flg
X30 Chronic_Osteo_flg
X32 Chronic_DIA_flg
X33 Chronic_DEP_flg
X34 Chronic_CPD_flg
X35 Chronic_CHF_flg
X36 Chronic_CAD_flg
X37 Chronic_AST_flg
X40 prof_visits
X41 fop_visits
X42 er_visits
X43 inp_admits
X44 tmc_amt
X47 chronic
X48 tds
X49 wellness
X50 lifestyle
32 Variables
Average Distance
Healthy 1
Abnormal 10.7
Avg. distance
There is a good separation between healthy
group and abnormals is higher with optimal set of
variables. Higher separation is desirable.
10
It is important to note that individual drivers are not the same even though they are subset of
population drivers This underscores the importance of individualized analysis
Root Cause Analysis With Contribution Ratios – Distance Sensitivity Table
ANALYSIS OF DIABETES CARE GAPS
X1 X2 X3 X5 X7 X13 X14 X16 X17 X19 X20 X22 X24 X25 X26 X27 X28 X30 X32 X33 X34 X35 X36 X37 X40 X41 X42 X43 X44 X47 X48 X49 X50
INDIV_
ENT_ID age gender_cdNew_Raceprospective_risk_numTDS_lbp_flgTDS_cad_flgTDS_osteoknee_flgTDS_osteohip_flgTDS_bngnutrn_flgWellness_prvntn_flgWellness_hlthyeat_flgWellness_hyprlpdm_flgLifestyle_Weight_flgLifestyle_Tobacco_flgLifestyle_Stress_flgChronic_Weight_flgChronic_Osteo_flgChronic_DIA_flgChronic_DEP_flgChronic_CPD_flgChronic_CHF_flgChronic_CAD_flgChronic_AST_flgprof_visitsfop_visitser_visitsinp_admitstmc_amtchronictds wellnesslifestyle
A1 31496 11.163 0.558 62.215 0.279 0.421 0.523 0.686 0.475 0.247 0.260 0.467 0.178 0.188 0.503 1.664 0.031 0.427 0.089 0.176 0.286 0.545 0.017 0.646 0.454 0.092 0.589 0.549 1.553 0.111 0.000 13.797 0.007
A2 57537 0.010 10.003 7.073 0.219 0.426 0.250 0.278 0.288 0.218 0.315 0.406 0.102 0.266 0.424 28.467 0.351 0.636 17.614 0.205 0.377 0.394 0.192 0.355 0.018 0.062 0.314 0.559 0.342 27.969 0.831 0.488 0.509
A3 59761 0.477 0.702 0.274 0.222 0.266 0.110 0.467 0.108 0.084 0.124 0.015 0.049 0.036 0.072 0.309 0.184 0.048 0.018 0.005 0.128 0.160 0.029 0.035 94.044 0.286 0.868 0.437 0.210 0.050 0.037 0.013 0.089
A4 69444 39.622 4.683 0.472 0.269 0.382 0.243 0.404 0.226 0.213 0.272 0.292 0.279 0.089 0.004 16.018 1.883 0.247 8.282 1.272 0.281 0.252 0.290 0.288 0.029 0.488 0.179 0.245 1.797 12.056 0.011 8.561 0.222
A5 76105 0.027 0.000 0.069 0.064 0.118 0.618 0.029 0.071 0.148 0.004 96.553 0.060 0.043 0.044 0.034 0.045 0.055 0.110 0.068 0.121 0.103 0.044 0.060 0.054 0.029 0.049 0.095 0.079 0.111 0.068 0.037 0.042
A6 79558 0.587 0.013 0.164 0.086 1.436 1.428 0.148 0.794 0.380 9.478 2.962 34.663 0.122 4.673 0.275 0.000 0.738 2.359 8.035 1.564 0.208 10.126 0.970 13.188 0.001 2.186 0.014 1.899 0.032 0.015 0.996 0.110
A7 98698 0.012 0.003 0.210 0.082 0.080 0.103 0.087 0.084 0.085 0.077 0.079 0.042 0.013 0.043 0.319 96.897 0.183 0.050 0.270 0.081 0.074 0.055 0.079 0.071 0.048 0.054 0.080 0.116 0.001 0.162 0.394 0.030
A8 103039 0.865 10.435 5.760 0.201 0.348 0.265 0.320 0.316 0.311 0.333 0.267 0.028 0.346 0.330 28.727 0.266 1.128 16.882 0.242 0.411 0.456 0.330 0.434 0.005 0.026 0.446 0.353 0.279 27.922 1.026 0.317 0.591
A9 107666 0.164 0.114 0.851 0.025 0.204 0.160 2.043 0.138 0.045 3.686 0.063 0.124 0.015 0.013 0.150 0.240 2.954 0.463 0.118 0.081 0.132 0.189 0.173 0.106 0.056 0.074 85.473 1.747 0.043 0.074 0.089 0.137
A10 116634 11.512 9.712 15.077 0.680 0.854 0.287 0.095 0.831 0.418 0.540 0.044 0.006 16.470 0.212 0.412 0.098 0.000 1.455 0.006 1.062 0.504 0.302 0.305 0.419 0.560 0.405 0.721 0.506 0.592 0.439 0.173 12.764
A11 121729 3.399 7.195 0.400 0.280 0.061 1.324 0.000 0.931 1.497 0.443 0.291 1.038 0.382 0.364 5.405 2.028 0.200 0.011 0.512 1.200 0.633 0.147 1.065 15.852 23.930 1.043 26.131 0.372 0.655 0.511 0.899 1.729
A12 138354 87.517 0.104 3.747 0.063 0.340 0.119 0.191 0.077 0.090 0.221 0.218 0.265 0.033 0.093 0.770 0.269 0.103 0.000 0.235 0.169 0.226 0.166 0.201 0.001 0.388 0.094 0.175 0.799 0.077 2.170 0.850 0.165
A13 139632 4.174 34.331 44.792 0.311 0.290 0.417 0.503 0.321 0.167 0.024 0.152 0.046 0.165 0.243 1.260 0.148 0.007 0.005 0.014 0.873 0.368 0.066 0.210 0.042 0.315 0.290 0.529 1.150 0.533 0.001 7.591 0.003
A14 146940 0.433 0.662 2.188 0.110 0.175 0.236 0.208 0.118 0.098 0.065 0.031 0.428 0.081 0.197 0.172 0.164 1.318 0.705 90.678 0.209 0.143 0.142 0.200 0.009 0.145 0.102 0.164 0.076 0.208 0.002 0.375 0.032
A15 186221 0.015 0.007 0.041 0.073 0.031 0.078 1.843 83.184 0.161 0.168 0.252 0.040 0.153 0.684 2.538 0.640 0.864 0.186 0.176 0.135 0.129 3.768 0.184 0.174 0.975 0.056 0.029 0.080 0.102 3.080 0.009 0.106
A16 194970 22.963 10.629 10.758 0.584 1.228 0.816 0.434 0.266 0.135 0.246 0.499 0.334 0.230 0.492 1.173 0.352 0.244 0.036 0.053 0.793 0.361 0.438 0.549 1.204 1.233 32.901 3.751 0.828 0.497 0.016 5.342 0.048
A17 202042 0.003 0.142 0.768 0.119 0.202 0.030 0.596 0.289 0.170 0.244 0.030 0.063 0.208 0.046 0.517 0.061 0.098 0.114 0.003 0.054 0.201 0.024 0.053 93.823 0.003 0.506 0.656 0.064 0.271 0.284 0.137 0.128
A18 219945 14.768 0.039 0.027 4.276 0.533 0.596 0.681 0.389 0.342 0.647 0.040 0.171 0.732 0.001 25.532 0.072 0.541 15.653 1.249 0.332 0.293 0.723 0.176 0.035 0.275 0.016 0.064 1.369 29.432 0.437 0.025 0.508
A19 228624 31.240 10.411 0.971 0.400 0.631 0.509 0.934 0.620 0.320 0.242 0.623 0.320 0.051 0.207 3.977 0.004 0.326 0.535 0.061 0.319 0.309 0.135 0.791 0.316 0.183 0.203 0.278 0.856 0.196 0.013 41.896 0.127
A20 251635 0.368 2.475 0.048 0.249 0.545 0.148 0.048 0.494 0.468 0.645 0.076 0.004 30.344 0.034 0.097 0.000 0.049 1.122 0.067 0.419 0.225 0.014 0.095 0.323 0.000 0.132 0.544 0.453 0.248 0.113 0.004 25.467
A21 252269 0.018 0.001 4.868 0.141 0.176 0.006 0.246 0.013 0.516 0.100 0.193 0.093 0.024 0.334 0.214 0.015 0.123 0.059 0.138 0.415 0.449 90.428 0.337 0.013 0.038 0.216 0.226 0.213 0.131 0.072 0.039 0.052
A22 259781 0.117 5.468 0.008 0.115 0.183 0.129 0.134 0.122 0.141 0.067 0.130 0.133 0.765 0.106 15.334 24.530 0.164 35.850 0.110 0.131 0.114 0.200 0.123 0.150 0.127 0.101 0.140 0.137 12.228 0.198 2.289 0.003
A23 265805 0.532 1.112 0.201 0.208 0.289 0.148 0.094 0.285 0.262 0.004 0.008 0.938 2.980 0.280 0.362 1.154 1.937 0.038 72.871 0.450 0.493 0.147 0.113 0.003 0.133 0.080 0.197 0.015 0.174 2.278 3.405 1.957
A24 274688 5.647 0.052 2.554 0.838 0.107 56.843 0.000 2.669 0.298 0.021 0.462 0.772 0.749 0.118 1.361 4.978 1.273 0.290 0.055 0.421 0.052 1.372 0.256 3.159 2.360 0.266 0.406 0.279 2.011 6.473 2.715 1.095
A25 277406 0.117 0.010 0.057 0.057 0.050 0.144 0.588 0.271 0.222 9.801 0.001 0.029 0.070 0.091 0.019 82.716 0.107 0.260 0.209 0.207 0.219 0.119 0.012 0.018 0.062 2.478 0.260 0.241 0.379 0.145 0.075 0.000
11
T R A N S A C T I O N A L N E T P R O M O T I O N A L S C O R E ( t N P S ) S T U D Y
S u m m a r y
Approach
KDCTM
Analysis
1394 variables
Target & Off-Target
observations
16 Variables Personalized-
Results
Impact tables
What has been done? Personalized Results (Sample Output)
KDCTM Methodology executed for tNPS data with a target
group of more than 38000 people and 17225 off-target
people
Distances of all 17225 off-target people
Overall population drivers of distance
Personalized distance drivers
Distances of 17225 off-target people
Contribution of each variable to all 17225 off-target
people
Impact analysis (based on individual variables and
combinations of variables)
Entire analysis done in less than 2 days
KDC™ has ability to add any new variables
KDC™ has ability to add customers even if they have not
taken tNPS survey
12
ACTIONABLE
ACTIONABLE INSIGHTS
KANRI Purpose-Driven Insights 14
Deliver data that explains performance gap root causes
Drive insights down to the individual customer/participant level
Simple, understandable output in terms on distance from target and root
cause drivers in percentages
Straight forward methodology that does not require a team of data
scientists to define and kick off
w w w. k a n r i - i n s i g h t s . c o m