How artificial intelligence (AI) can help maximize customer intelligence ROI
0 likes107 views
This document discusses how artificial intelligence can help maximize customer intelligence ROI. It provides an overview of customer intelligence and discusses how AI is enabling more data, more predictive analytics, and more automation in customer intelligence. Specific use cases discussed include churn prediction and next best action recommendations for a large European bank, product recommendations for a B2B software vendor, and large-scale consumer insights for a global CPG company. The document concludes with best practices for AI projects, emphasizing defining processes, including business stakeholders, exploring and iterating on features, combining supervised and unsupervised machine learning, and robust deployment.
![Large Scale AI-Enabled Consumer Insights
CRM, Social
media, Web &
mobile analytics,
Partner Data,
Customer Service
Data
Are people talking about
my product ?
What’s the next big flavour
in Ice Cream ?
What do people say when
they talk about tea ?
INTERNAL DATA TEAM
(200+)
Data Scientists
Analysts
Marketers
Sentiment Analysis
Topic Extraction
Forecasting
Trend Spotting
Scoring
Dashboard
Insight & Data
Data Products
“In 2018 we generated savings in BMI
[Brand and Marketing Investment] of
over €500 million...
Improvements to measurement and
verification of digital audiences ensure
we maximise value in digital advertising
alongside improvements in the
measurement of influencer follower
data.”](https://image.slidesharecdn.com/howartificialintelligenceaicanhelpmaximizecustomerintelligenceroi-latestdataiku1-200824095455/85/How-artificial-intelligence-AI-can-help-maximize-customer-intelligence-ROI-19-320.jpg)
How artificial intelligence (AI) can help maximize customer intelligence ROI
- 1. How Artificial Intelligence (AI) Can Help Maximize Customer Intelligence ROI By Vincent De Stoecklin Customer Success Director at Dataiku
- 2. AGENDA AI for Customer Intelligence ❏ Defining Customer Intelligence ❏ Some Historical Perspective ❏ The Big Data and AI (r)evolution ❏ Key AI Use Cases In Customer Intelligence ❏ Real Life Customer Use Cases ❏ From Idea To Impact – Challenges And Key Success Factors
- 3. CUSTOMER INTELLIGENCE IN THE ERA OF AI
- 4. Introductory Overview Customer intelligence (CI) is the process of gathering and analyzing information regarding customers, and their details and activities, to build deeper and more effective customer relationships and improve decision-making by vendors -- Wikipédia Chart by Finextra
- 5. ROI Of Customer Intelligence Programs Customer Acquisition Key ROI metric : Cost of acquisition > How much does 1$ of revenue (profit) from a new customer cost me? Customer Loyalty Key ROI Metric: Weighted lifetime value (rebuy, upsell, cross-sell) > How much does an additional 1$ of revenue (profit) from an existing customer cost me? Chart by Finextra
- 6. Should You Focus On Acquisition Or Loyalty? UPSELL CROSS SELL IT IS MUCH MORE COST EFFECTIVE TO SELL TO AN EXISTING CLIENT Benchmarks on upsell and cross-sell weight in companies • Subscription business – approx 16% of revenue (source: insightsquared) • Non-subscription business – between 0,5% and 4% (source: getelastic) Source: Groove
- 7. Priority Use cases in Customer Intelligence Business Value Complexity Churn Prediction Omnichannel Product Reco. Retargeting Customer ServiceReferral Programs Marketing Attribution Dynamic Pricing Next Best Action Compliance & Regulatory Cost Reduction Branding Revenue Increase Customer Satisfaction Overall Value Availability of data Complexity of data Complexity of modelling Complexity of deployment Overall Complexity Cost / Benefit Analysis – Use a Quick Proxy
- 8. SO WHAT’S NEW WITH AI? AI or not AI…
- 9. Some Historical Perspective ? $$ In the early 90’s technology vendors working with large retailers found out a strange correlation…
- 10. Customer Intelligence at Scale MORE DATA MORE PREDICTIVEMORE AUTOMATION What changed in the world of Big Data / Machine Learning / AI? Customer intelligence (CI) is the process of gathering and analyzing information regarding customers, and their details and activities, to build deeper and more effective customer relationships and improve decision-making by vendors -- Wikipédia
- 11. More Data: The 360 Customer View • New dimensions available for customer intelligence • Combining internal + external data • Transactional + behavioural + socio-demographic • More granularity in the available data (faster refresh, identifiable individuals)
- 12. More Predictive – The Rise of ML/AI • No preconceived rules or correlations • Less need for interpretability • Aligns segments to business objectives • Dynamic evolution of segments
- 13. More Automation – 1:1 Marketing • High degree of automation and granularity for outcomes of CI programs • Marketing 1:1: hyper-personnalisation of segments and offers • Automation of interaction scenarios for customer loyalty and satisfaction
- 15. Churn And Next-best Action For A Large European Bank Automated Integrated to the IS FlexiblePerformantIndustrial In line with Business Strategy INITIAL REQUIREMENT Increase customer retention for existing customers by anticipating churn and recommending relevant offers and services via marketing channels (email, website, in-branch…) ● Each client can should have a score for churn risk and have multiple possible “actions” offered, ranked by relevance (i.e probability / value) ● Complex business rules to be added into the system, and should be overridable by humans Self-LearningRelevant Dynamic
- 16. Project Ecosystem and Outcome CHURN PREDICTION AND NEXT BEST ACTION SCORING 15M CLIENTS EVERY 3 MONTHS SERVICES CATALOGAPP / WEBSITE NAV LIFE EVENTS CUSTOMER PROFILE Recommender system is completely integrated with the activation engine • Connected to ADOBE CAMPAIGN email marketing • Connected to bank agents’ front end in-branch insight generation • « Human Overridable » based on current strategy and corporate guidelines • IMPACT: LIFT x3, RETENTION +12%
- 17. Product Recommendation For A B2B Software Vendor • Sales software vendor in the US • More than 4000 clients (Dell, HP, EMC…) • Wanted to propose a built-in recommendation engine to help its clients generate more revenue Real time recommendation engine
- 18. Next Product To Buy In Real-time • Real time predictions with explainability • Retraining of models specific to each client • Champion / challenger mode • Average basket +7%
- 19. Large Scale AI-Enabled Consumer Insights CRM, Social media, Web & mobile analytics, Partner Data, Customer Service Data Are people talking about my product ? What’s the next big flavour in Ice Cream ? What do people say when they talk about tea ? INTERNAL DATA TEAM (200+) Data Scientists Analysts Marketers Sentiment Analysis Topic Extraction Forecasting Trend Spotting Scoring Dashboard Insight & Data Data Products “In 2018 we generated savings in BMI [Brand and Marketing Investment] of over €500 million... Improvements to measurement and verification of digital audiences ensure we maximise value in digital advertising alongside improvements in the measurement of influencer follower data.”
- 20. BEST PRACTICES FOR AI PROJECTS
- 21. Quick Insights On AI Project Lifecycle 1. Define process around project scoping & prioritisation 2. Think about the outcome and include business from A to Z 3. Explore and iterate to create relevant features 4. Combine supervised and unsupervised approaches 5. Deploy and operationalize your data flows in a robust and governed way Scope Scoping Scope Scope
- 22. Q&A
- 23. Thank You