Share
Embed
Cross-Channel Analytics and Reporting
Andreas Volmer, Voxeo Sr Product Manager
Voxeo
October 14, 2011
Transcript
21 slides
-
Reporting Challenges • No insight into customer experience • Gap
between implementation, documentation, and reports • Business don’t get the metrics they want • No consolidated reporting across channels -
Application Lifecycle Management VoiceObjects Analyzer Grammar IDE Debugger, Trace Viewer,
Phone-less testing, Load Testing Project Documen- tation Storyboard Manager Desktop for Eclipse Control Center -
Design, Documentation, Reporting Call Flow Implementa.on (Desktop for Eclipse)
Documenta.on (PDF) Repor.ng and Analy.cs (VO Analyzer) -
Design, Documentation, Reporting Call Flow Implementa.on (Desktop for Eclipse)
Documenta.on (PDF) Repor.ng and Analy.cs (VO Analyzer) -
Out-‐of-‐the-‐box Data Capture Dimensional Data Model, Real .me data
Ready for Integrates in Data Warehouse CTI Data IVR Data CRM Data Infostore VoiceObjects Analyzer (Business Objects, Cognos, MicroStrategy) Analysis and Repor.ng Based on standard BI tools 60+ predefined Reports VoiceObjects Infostore and Analyzer -
System Usage & Performance Server Stats Session Details Application &
Grammar Tuning ASR Performance Utterance Statistics Utterance Recordings Customer Behavior Intelligence Dominant Path Analysis Personalization Stats Business Tasks VoiceObjects Server Data Capture Backend Stats -
Drill through to Utterance Analysis for lGet Credit Card Typez
… Recognition Performance Analysis -
Grouping by Interpretation (lSlotsz) All utterances recognized as lVisaz Full
phrases with filling words (lI havez, lpleasez) 5.25% of callers used DTMF input, often triggered by NoMatch events Grammar Tuning: Utterance Analysis -
Back-end Performance • Identify back-end performance issues • Real-time monitoring
and post-mortem analysis • Define performance metrics and alerts – Take action before errors occur in live system -
System Usage & Performance Server Stats Session Details Application &
Grammar Tuning ASR Performance Utterance Statistics Utterance Recordings Customer Behavior Intelligence Dominant Path Analysis Personalization Stats Business Tasks VoiceObjects Server Data Capture Backend Stats -
Question: How does call flow personalization work for our callers?
At call termination, 78.6% of callers had authenticated Call duration is shorter for Platinum users Application success analysis by Layer states (e.g. Customer segments, …) Personalization Reporting -
Reporting on Business Tasks" • Questions to be answered: How
many callers tried to accomplish a certain task? How many were successful? What steps did it take to succeed? How long did it take them? How well did speech recognition work for this task? Why did task completion fail? What needs to be fixed? -
… and there were recognition problems and hangups while entering
new credit card data lUpdate Expiration Datez was very successful … … but 82% actively aborted the lOrderz transaction … Business-level statistics on Task completion rates Cross-channel Reporting Business Tasks: Completion Rates -
… and there were recognition problems and hangups while entering
new credit card data Business Tasks: Completion Rates -
Show only Calls where the task lEnter New Credit Cardz
failed due to a recognition failure Step 1: Define Filter Criteria … only Calls that terminated in a given Input State or Module … only Calls with certain customer segments Individual Call Analysis -
Step 3: Analyze Session Inspect dialog steps associated with the
task Caller hung up after 2 failed speech recognitions attempts Listen to utterance recordings Individual Call Analysis -
VoiceObjects Analyzer Analyer Usage Scenarios Analytics Interactive Web frontend Slice
& dice, Drill-down Application Tuning Mobile Access Management-level KPI and trend analysis reports Alerts / Notifications Metrics definition Alerts via SNMP, email, ... Dashboards Performance Overview Service Usage Monitoring Report Generation Automatic report creation (scheduled daily / weekly) Send as PDF or HTML Data Warehouse Service usage and success analysis by customer segments