Customer Analytics Strategies for Success John Keyes Sound

Customer Analytics: Strategies for Success John Keyes Sound. Bite Communications

Agenda • Overview of analytics • Contact center analysis • Outbound analytics and optimization 2

• Performance Analytics – Reportsof that. Analytics show past Overview performance and trends • Predictive Analytics – Models that make predictions / forecasts Help to answer the following questions: • How has my contact center been performing? • Should I contact this account? • What has been the best day and time to reach call recipients? • When should I change my treatment and/or script strategy? • Have recipients been responding across all types, scores, balances, and ages? • If so, what should my method of contact be? • When and how often should this account be contacted? • What is the probability of this account responding? • How much should I spend to try to get this account to respond? 3

3 levels of treatment Level Description Statistical Examples Data Mining • Performance Analytics Portfolio Level Treat the entire population in the portfolio the same. • Segment Level Create sub-groupings of the population and treat each sub-group the same. • Record Level Treat each individual in the population as a unique entity. • Regression Models • What-if Simulations Clustering • Segmentation 4

Analytics Progression Category Description Examples Measurement and Reporting Tracking and summarizing campaign results using predetermined or ad-hoc metrics. • Campaign Reports • A/B Testing and Measurement Strategic Data Analysis Monitoring and analysis of trends in data to produce conclusions with significant business meaning and strategic inputs. • Performance Analytics • Trending Analysis • Seasonality Analysis Heuristic / Data Driven Segmentation / Scoring Organizing customers into meaningful groups for business management or differential treatments based on business rules / decision trees / clustering techniques. • Balance-score Based Segments/Scoring • Multi-attributes Data Driven Segments/Scoring Predictive Modeling/Scoring Applying statistical modeling techniques to predict and maximize individual customer behavior. • Likelihood to Pay • Recovery Score, Action Score Simulated Optimization Using optimization algorithm to assign customer treatments that satisfy defined business objectives and resource constraints. • What-if Scenario Simulations 5

Examples: Performance Analytics Contact Center Performance Analytics • • Hold times by hunt group (including time of day / day of week) Recipient willingness to hold Hold and ring time trends Agent talk time distributions across campaigns Call Recipient Performance Analytics • Call pass & device escalation effectiveness • Best time of day / day of week • Responses across locations (region, state, zip code, area code, time zone, etc) • Script hang-up graphs • Responses by recipient characteristics (score, balance, etc. ) • Test and Control performance results (A/B Testing) 6

Integration with analytics partners Analytical Data Client Database Partner: Should this account be called? “Yes” • Calling Strategy • Optimization • Analytics “No” Holdout or Other Treatment 7

Contact Center Analytics 8

Contact Center Service Level Analytics Willingness to Hold 9

Analyzing the Components of Wait Time 4 Seconds 18 Seconds 20 Seconds 8 Seconds 10 Seconds 12 Seconds 50 Seconds 22 Seconds 10

Time of Day in the Contact Center “My average service level is 18 seconds so I don’t need to make any improvements to my contact center. ” Tuesday Contact Center Analytics Time of Day Busy % No Answer % Success % Hold Time 9: 00 A. M. 18% 1% 62% 48 seconds 10: 00 A. M. 6% 0% 78% 22 seconds 11: 00 A. M. 2% 3% 92% 16 seconds 12: 00 P. M. 3% 15% 72% 28 seconds 1: 00 P. M. 1% 2% 77% 16 seconds 2: 00 P. M. 1% 1% 87% 12 seconds 3: 00 P. M. 2% 0% 92% 10 seconds 4: 00 P. M. 3% 0% 92% 14 seconds 5: 00 P. M. 7% 1% 76% 21 seconds 6: 00 P. M. 11% 3% 68% 37 seconds 11

Skill Group Reporting 12

AVM / Contact Center Analysis 13

Skill Group / Right Party Talk Time Analytics 14

Outbound Analytics and Optimization 15

Calling Time Zone / Time Focus Analysis 16

Call Pass Effectiveness Analysis 1 Pass 25 – 30% Live% 50 – 55% AM Total Inbound: DCs: 30 Callbacks: 22 3 Pass 65 – 75% Live% 15 – 25% AM Total Inbound: DCs: 75 Callbacks: 10 17

Day of Week Analysis 18

Time of Day Analysis 19

Banding Results: By Balance 20

Banding Results: By Age 21

Banding Results: By Score 22

Script Effectiveness 23

Device Escalation Effectiveness Analysis Work, Home, Cell 1. 88% Promise % Home, Cell 2. 18% Promise % 24

Multi-lingual Scripting Options Add Spanish Option 25

Geographical Analysis-Voice Dialects Country Time Zone Region State Area Code 26

Dashboard / KPI Management 27

Summary • • • Start with contact center performance analytics Outbound performance analytics KPIs and measurement A/B testing Predictive Analytics • Don’t just look at averages • Drill to the lowest level • Measure, Measure! 28

Thank you! John Keyes Sound. Bite Communications jkeyes@soundbite. com 781 -359 -2236 29