Making Marketing Smarter with Analytics Prof. Francisco N. de los Reyes Analytics Advisor Thakral One Measurement and Data Science University of the Philippines School of Statistics
Big Data Demographics Transactional Data Customer Value Lifestyle Usage Behavior Expectations Needs/Wants Psychographics Personality Sentiments Egonets Touchpoints Channels Risk Flags Remote Sensing Data and may I add, VOICE! VARIETY VOLUME VELOCITY
Making Marketing Smarter Harness the synergy of data, statistical tools and social media.
Why Smarter? Complex processes (i.e. major purchase decision) are viewed in various angles through a variety of platforms. Cutting-edge computing tools are utilized to demystify the customers paradigm. Identify real points in time to anticipate and redirect decision, arrest objections and infer real needs. Connect knowledge of customer to external shocks (environmental, political, temporal) as they happen. Personalize messages, offers and rewards at a precise time thus delighting the customers and transforming them into brand heralds.
P(Next Purchase not Later than Day T) From Classical to Innovative Advances in statistical science enhance predictive and prescriptive capacities. For example, churn models are made better by predicting the time-for-next purchase rather than setting a fixed horizon and predicting the churn event. 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Day T 2 4 6 Time-Related Purchase Probabilities 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 Mid Value Churn-Prone High Value Mid Value Regular Objective: Determine when a customer is in his/her highest likelihood of making a purchase. Traditional approach answers the question: In the next 3 months, who is likely to leave me? Innovative approach answers the question: When is Customer Joe likely to make another purchase? In 10 days? 30 days? 60? 90? Impact: Precise Timing of Offers If propensity of purchase is slow, marketing can accelerate next purchase. If campaign timing is set on an earlier date, the latter churners in a 90-day period may not pay attention. If campaign timing is set on a latter day, the likely-churners may have already churned earlier. Campaign is thus less effective.
From Classical to Innovative (Linked Segmentation) Patient Segmentation Doctor Segments One segmentation leads to another segmentation that targets loyalty.
Web & Social Media
Web Engagement Expansion Rate Click Impression Dwell Time CTR Video Duration Click to Conversion Interaction Mobile data Interaction time Widget spread Dwell Rate Play Rate
Improving Customer Communication (Track Connectedness) Offers, rewards and communication may halo to a specific customer s inferred social circle.
Data Challenges Desire: Comprehensive Data Warehouse for Analytics Challenges: Data in Silos - need to integrate under one Single Customer View Defining Metrics - from seeming noise to signals Quality of Raw Data - accuracy, believability, objectivity, reputation, value-added, completeness, relevance, appropriateness in volume, interpretability and ease of understanding, accessibility, and security
Trends New data sources - Wearable devices, personal recognition technology, culture (and subculture) blogs, live media audience monitoring, environmental scans and crowd sourcing apps on weather conditions, traffic situations, special events, etc. Data Science - Massive framework for storage and processing (e.g. Hadoop); multivariate statistical methods, advanced time series analysis and forecasting, high dimensional data visualization, text networks, random forests, latent Dirichlet allocation; subject matter researches (e.g. psychology and media). Visualization - Interactive 2D and 3D displays, query layers, overlay maps, hyperspectral satellite images, principal component maps, heat (kernel) maps. Best Practices for Enhance Customer Experience - Relevant offers based on inferred need, behavior and social circle; personalized rewards based on customer s own consumption pattern; personalized service.
Elevate Analytics Efficient Data-Integration Comprehensive Customer Views Majority of data are highly granular. Ensure usefulness of data by careful integration into a single customer view for efficient translation into an analytics base table. Updated Tools Abreast with Data Science Techniques Big data is analyzed using novel approaches far advanced from classical techniques. Simulation, resampling, back fitting and neural network approaches are now quite involved. Avoid black-box approaches. Protect Customer Privacy Being Increasingly Social Facebook profiles, applications, Twitter feeds, YouTube uploads A good proportion of Cannes winners used social networking as an element of their marketing campaign. Presence in Multiple Screens and Channels Applications, photos, Facebook profile updates, games Customers are savvy to digital. access of the internet via PC, tablets and mobile phones are more intense than ever.
Thakral One Value Proposition BUSINESS EMPOWERMENT THROUGH ANALYTICS Maximizing Existing Assets Establishing a Scalable Foundation Operationalizing Analytics Knowledge Transfer & Enablement Management & Risk Consulting Our People Our Experience Our Access to Best Technologies
Thank You francisco.delosreyes@thakralone.com