INSIGHTS ANALYTICS INNOVATIONS Data Science & Big Data Practice Customer Intelligence - 360 Insight Amplify customer insight by integrating enterprise data with external data
Customer Intelligence 360 Insight -Introduction Attract Who are my customers and how to attract them?. Customer intelligence is the process of knowing a customer better. DSI helps organizations to know about their customer by amplifying customer insight through integration of internal data (past purchase behaviour, response to past campaigns, complains, etc.) with external data (social graph, demographic data, etc.). The prolific growth in Customer base, with the Personalize What personalized offerings can I make for max revenue? Reach What is my strategic customer center of gravity? Location, Macro- Economic Marketing, Inventory CRM Enhanced 360- Degree View Transactions, Sales Social Media Measure What do they purchase and how? Monitor What is their socio-web activity? advent of new tools and techniques, drives this need to build an intelligent 360 degree framework for unified view of your customers and their digital Retain What steps should I take to retain them? Campaign, Communication Call Center, Social Web Risk, Models Analyze/Evaluate How do I segment and score customers? footprints. Engage How am I engaging with my customers?
DSI Customer Intelligence Offerings Customer Acquisition, Service Delivery Customer Acquisition Customer Intelligence DSI Customer Intelligence solutions help business leaders put their data to action, discovering opportunities to create short-term impact and long-term enterprise value Customer Service Delivery DSI is using advanced data science to capture a 360 view of a customer s interactions and experiences across channels in real time. This Customer Intelligence offering allows you to respond intelligently when a customer engages your organization at any touchpoint and helps you to determine the appropriate go-to market strategy and offer personalized product/services. Increase New Customer Base Winning back Lost Customer Revenue Increase Customer Satisfaction Customer Life Time Value Up-Sell Cross-Sell Voice of Customer
Customer Acquisition Business Leaders using Customer Intelligence to increase acquisition
Increase Customer Base DSI s unique approach of Customer Analytics transform raw data into actionable insight so that business can effectively understand the need of the customer and explore innovative ways to improve customer acquisition. Our advanced analytical model will enhance customer acquisition, optimize pricing of services across the consumer lifecycle, and maximize profit by the knowledge of when/where product promotions will be most effective. Methodology Utilizingsegmentation techniques for determiningthe most appropriate group of companies to target a message, product or service. Our data science solution employ identifying which digital marketing tactics plan to use as an acquisition technique. Ad optimizationprovides targeting or tailoringadvertising (placement and message) to attract a specific client. Business Outcomes Maximize revenue through Increased customer acquisition. Minimize the operational costs by targeting right customer that are most likely to respond. Identification of products, services and features that customers care about most. Optimize budgets and communication channel to ensure fund is allocated to the best performingprograms, and report on performance. INSIGHTS. ANALYTICS. INNOVATIONS.
Win back Lost Customers Customer Segment Value Many businesses are singularly focused on finding as many new customers as possible. That might seem like the most effective strategy, but recovering customers who have switched to another vendor is actually a lot easier than acquiring someone new. Moreover, existing customers are a potential source of easy-to-close referral sales. Even a single lost customer can have a ripple effect that increases cost of sales. DSI can help to reconnect with the customers switched to another company or service provider Methodology DSI provides analytical solution for Identification of lost customer based on the company historical data base. Categorizing the customers who are likely to comeback considering their reason of leaving, purchase behavior, demographic and socio-economic profile. Our advance analytical solution determine most potential and profitable customers based on their customer lifetime value (CLTV). Design product and services to match the need of different customer segments Utilize campaign analytics and channel optimization for exploring innovative ways to improve customer targeting with right communication channel. 0 Business Outcomes 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Low priority 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Probability of Return High priority Bubble size represents # of customers in a segment Insights for winning back selective churned customer back into customer base. Identification of products, services and features that led to increased customer satisfaction
Customer Service Delivery
Customer Service Management -360 View Customer interaction is an ever evolving process that becomes more and more complex as time passes. DSI s Customer Interaction analytics helps business to gain insight into how market responded to campaign,what typical products/ service concerns are,how each geography/customer is different than others etc. Methodology Utilizing speech,text,and other raw data to create a holistic view of Customer Interaction. Create actionable insights from quality assessments,sentiment assessments, Customer interaction intentions and others. Business Outcomes Scorecard identifying customer concerns,visuals highlighting trends and patterns and providing early warning predictive insights. Predictive insights for Customer churning,targeted up-sell and cross-sell activities. Further Build KPIs that are tuned to customer value to identify and group customers by net value so your most important customers are given a differentiated and personalized experience that builds profitable brand loyalty Enhanced 360 Degree view of Customer to deliver superior customer experience. Roadmap for enhanced customer experience at different point of service. INSIGHTS. ANALYTICS. INNOVATIONS.
CLTV Customer Lifetime Value Customer Lifetime Value is prediction of revenue / profit attributed to entire future relationship with a customer. DSI developed prediction model can have varying levels of sophistication and accuracy, ranging from a crude heuristic to the use of complex predictive analytics techniques. Methodology Observe various individual-level buying patterns from the past - find the various customer stories in the data set. Forecasting of future revenues based on estimation about future products purchased and price paid High Medium Highest Priority CLV-based segmentation is combined with a Share of Wallet (SOW) model to identify "high CLV but low SOW" customers to focus marketing resources for this customer segment Low Lowest Priority Business Outcomes Quantifies likely future business transaction with different customer segment High Medium SOW Low Provides ROI based decision support for marketing activity Optimize marketing spending for maximum value rather than minimum cost Cost to Serve CLTV INSIGHTS. ANALYTICS. INNOVATIONS.
Transaction Value Up sell & Cross sell Digitization of the financial industry and a rising competitive landscape demands the need to have a holistic view of your customer. Adapting to your changing customer requirements needs an ability to have the predictive capability for identifying opportunities of cross selling and up selling right products to the right customerand at the right time. Methodology Analyse customer segments and their buying patterns. Cluster customer segments based on their purchase and product usage pattern. Apply Data Science solution to identify product basket mix across customer clusters. Use custom scoring models with integrated business rules for recommendation to customers. Up Sell Window Low Up Sell potential Average Value Curve for population segment Up Sell potential high Business Outcomes Product Segments Linked to Customer Segments for Cross-Sell Identification Increase incremental cross sell, up sell, and deep sell revenue Isolates and eliminates the negative effects of marketing Reduces marketing spend while generating more revenue Increased Customer Satisfaction Index. INSIGHTS. ANALYTICS. INNOVATIONS.
Transforming your Data Chromosome 1 1 0 1 0 1
40. 00 35. 00 30. 00 25. 00 20. 00 15. 00 10. 00 5.0 0 0.0 0 200.00 180.00 160.00 140.00 120.00 100.00 80. 00 60. 00 40. 00 20. 00 0.0 0 Case Study Upsell and Cross sell The client is a leading manufacturer and distributor of jewellery having 120+ franchisees. The client was experiencing stagnation in same-store sales growth over the last few years. The client introduced many new SKUs to provide more choices to customers. This action resulted into proliferation of SKUs and overall inventory at franchisee level. The client was forced to accept unprecedented return of material from franchises in the last 6 months. Few franchises left the client and stores were closed down. Business Questions Who are my profitable and loyal customers based on expected revenue in next couple of years from different customer segments Which type of products from existing 50000+ SKUs will be suitable for which segment? Can we predict next purchase behaviour (type of product, value of purchase, time of purchase, etc.) of the important customer segments Where do we see higher concentration of a specific customer segment? How can we rationalize product portfolio at store level based on concentration of specific customer segments? Value Propositions Illustrative Solutions Mapped purchase pattern (type of product, billing size, buying frequency, reason for purchase (gift, self) etc.) with socio economic profile (age, gender, education, income, asset ownership, location, etc.) of customers screened the socio economic profile of the people residing in the catchment area of the existing stores Estimate purchase value by different customer segments Establish the relationship between customer segment and likely purchase pattern based on last few purchase history Business Impact Same store sales went up by 12% YOY in the following 6 months in comparison to average 4% growth in the preceding 2 years. Inventory reduced from 123 days of sale (average of preceding 12 months) to 97 days of sale (average of following 6 months) Sales and Inventory Customer Segmentation Identify the profitable customer, promote right product, maximise sales, minimise customer defection CY-11 CY-12 CY-13 CY-14 Avg Same Store Sale Inventory (DOS) INSIGHTS. ANALYTICS. INNOVATIONS.
Case Study Customer Sentiment Analysis One of the largest multi national retail chains wanted to analyse customer perception about their stores and the brand through social media analytics. The major objective of the project was to determine location wise customer sentiment on important issues like customer service, product availability, product quality, etc. The company has its own Facebook page where customers express their sentiment about their store experience or their perception about brand promotions, advertisement, communication, etc. Business Questions How satisfied my customers are on different issues like customer service, product availability, price, quality, etc. What are the major issues customers are talking about How stores in different locations are managing customer perception Can we find out store wise issues (positive and negative) through social media perception analysis How my promotional campaigns are performing in terms of generating awareness and are they provoking positive or negative sentiment about the brand Value Propositions Illustrative Solutions Text corpus was mined using lexicons (industry specific, and English language) for text pre-processing, parts of speech tagging, word frequency and entropy analysis, sentiment and emotion analysis Topic Modelling keywords and phrases from the corpora was used in accordance with a SME knowledge graph to identify and model key topics and relationship between them. Topics helped in classification of customers to derive their interests and intent. Business Impact Identified location wise key issues. This helped the retailer to take action at store level Monthly trend analysis provided insight about change in perception amongst consumers by location Evaluated impact of marketing campaigns over a period of time Topic Identification Sentiment by Location Micro level issue based, location based sentiment analysis to correctly distinguish positive issues from the negative ones INSIGHTS. ANALYTICS. INNOVATIONS.
Investment Potential Case Study Winning back Lost Customers One of the leading asset management companies had millions of inactive retail customers in its database. These customers had invested with the company in the past but closed the investment at least 6 months back. The company started the process of reaching out to these customers but did not achieve success in this regard. The objective of the project was to go through the list of inactive customers and find out the customers who are more likely to respond positively to specific set of financial products. Business Questions How to segment the lost customers based on their likely investment potential. Why did these customers leave the company? Was it related more to macro economic environment or more to company s poor performance What is the right product to offer to different customer segments What is the best medium to communicate with these customers? How the choice of medium will vary because of change in the profile of the customers over time (especially applicable for those who left couple of years back) Value Propositions Targeting right product to the right customer through right channel of communication to improve conversion rate Illustrative Solutions Employed cutting-edge machine learning algorithms to segment customers based on demographic, economic, behavioural, and past transaction parameters. Analysed different types of product and matched them with customer s segment to ensure that each customer was targeted appropriately. Determined propensity score for each customer to join back the company Determined likely investment value for each customer segment Business Impact Improved conversion rate by more than 90% Overall investment from lost customers went up by more than 170% in the following six months 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Probability of Return Bubble size represents # of customers in a segment INSIGHTS. ANALYTICS. INNOVATIONS.
Case Study Campaign Management One of the leading electronics manufacturing companies periodically connects to its customers through direct mailing, in which it communicates about new products, promotional offers, etc. Response to these campaigns were not satisfactory for the company. The ink (for printer cartridge) division of the company was loosing market share slowly for the last couple of months. To turn the tide, the division came up with an ambitious plan to increase repeat purchase of their existing customers. It designed a new campaign, where a promotional coupon was sent with the email campaign to the customers. The Division wanted to promote only 3 products/services keeping the campaign short and relevant to their customers. Business Questions What are the 3 major products/services to be offered to different customer segments Which customer segments to be targeted for these campaigns What will be the coupon value which is likely to response rate for each customer segment Value Propositions INSIGHTS. ANALYTICS. INNOVATIONS. receive the max Illustrative Solutions Predictive analytics was used for campaign optimization, to identify the right customers for campaigning. Data mining algorithms were used to micro segment customers based on demographic profile, location, purchase pattern, response to previous campaigns, etc. Decision tree algorithm was created to turn customer insight into highly targeted segments to drive higher response rates and a more robust ROI. A/B testing was carried out to evaluate effectiveness of design on different customer segments. Business Impact The model improved overall response rate by 77% in comparison to past campaigns. 25% less customers (who are predicted least likely to respond were omitted) were contacted during the campaign, which helped in reducing cost. Customer Segmentation Identify the profitable customer, promote right product, maximise sales, minimise customer defection Response Rate (%) Cmpgn1 Cmpgn2 Cmpgn3 Cmpgn4 Cmpgn5 Newspaper Leaflets Banners Hoardings On-line
Case Study Marketing Mix Optimization A global bank was rolling out a limited period promotional offer on credit card to its customers. Additional reward points were offered to the customers based on card spent for a specific period. The bank was not happy with the ROI of similar campaigns in the past. The bank had traditionally used combination of certain media channels to promote campaigns. The bank was also tied by long term contract to provide certain amount of business to each media. The bank had already decided on newspaper, leaflets, display ads (banners/stickers), hoardings, and online ads as the media for promotion. But, it was not sure about how much it should spend on each of these medium to maximize the campaign ROI. Business Questions How much money should be spent on these mediums (newspapers/leaflets) to maximize the campaign ROI? Which are the profitable customer segments and what are their preferred medium of communication How to determine marginal benefit of incremental ad spend Value Propositions INSIGHTS. ANALYTICS. INNOVATIONS. What is the right sequence and frequency of exposure for each customer segment for each communication channel Illustrative Solutions DSI team analyzed media budget by each medium and corresponding response rate for the previous 5 campaigns. S-Curve was generated for each media based on response rate for the media and money spent on the media for the previous campaigns. The marketing mix optimization model was framed after going through each constraint in detail. The response rate was monitored post campaigning to evaluate effectiveness of the optimized media program. Business Impact The model suggested significant change in budget allocation to On-line and leafletit. On-line budget was increased significantly whereas the same for leaflet was reduced to half of the historical avg. The bank experienced more than 20% improvement in response rate for the campaign Response Rate (%) vs Media Spend ($) Response Rate (%) Increase customer response rate by optimizing media spend across multiple channels of communications
Data Solution Architecture Developer Tools and Adhoc Analysis 2 Big Data Lake 3 Text Mining Multivariate Hadoop Correlation Analysis Spark NoSQL Segmentation Machine Learning 4 Enterprise Unified Data View Customers Products Finance Risk Sales Operations Visualization, BI and Analytics Apps 1 Data Ingestion & Engineering (ETL, Engineering) Metadata, Semantic Social Media Website s Blogs Media Macro- Economic Enterprise Un-structured Data Enterprise Structured Data External Data 1 Data Engineering External/Internal Data Ingestion - Real time/batch Quality, transformation Extract Transform Load 2 Big Data Lake Parallel and clustered processing Structured & unstructured data Hadoop/Spark/NoSQL 3 Data Science Advanced Analytics, Machine Learning and statistical analysis Predictive Model, Recommendation INSIGHTS. ANALYTICS. INNOVATIONS. 4 Unified Data View Data Integration for Unified View Data warehouse for Viz, BI, Apps
INSIGHTS ANALYTICS INNOVATIONS INSIGHTS. ANALYTICS. INNOVATIONS.