© 2014 Cognizant BBVAOpen4U Innova Challenge SpendWise Genie Application Dec 1, 2014 SpendWise. Be wise © Cognizant Technology Solutions 2014. All rights.
Post on 23-Dec-2015
217 Views
Preview:
Transcript
© 2014 Cognizant
BBVAOpen4U Innova Challenge
SpendWise Genie ApplicationDec 1, 2014
SpendWise. Be wise
© Cognizant Technology Solutions 2014. All rights reserved. Cognizant owns all rights in all its trademarks, trade names or logos, Patents, Copyrights and any other intellectual property rights used in the presentation. Cognizant acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in the presentation. Except as expressly permitted, neither this presentation nor any part of it may be reproduced, stored in a retrieval system, transmitted or modified in any form or by any means, electronic, mechanical, printing, photocopying, recording or otherwise, without prior written permission of Cognizant Technology Solutions. Cognizant disclaims and makes no warranties or representations as to the accuracy, quality, reliability, suitability, completeness, usefulness of the presentation.
© 2014 Cognizant 2
What’s in store …• Background
• Introduction
• The Experience
• User Journey
• Key Insights
• “People Like Me” & “Offers for Me”
• Key Business Benefits Delivered
• How it Works
• Solution Architecture – Business & Data Flows
• Technical Architecture – Application Design Components
• Cloud Based Predictive Model Design
• Recommended Road Map
• Solution Evaluation Parameters
• Appendix
© 2014 Cognizant
Background
Banks face the challenge of leveraging large volumes of disparate data for increasing customer engagement (consumers & Bank affiliated merchants)
Merchants find it challenging to provide new and existing customers with target offers at the right place & time and increase sales
Customers need to manage personal finances, monitor spend and save money on their purchases through relevant offers
3
SpendWise. Be wise
Cognizant believes that Banks can address these questions by leveraging advanced Cloud Based Data Analytics and the Digital Banking–driven ecosystem ( including Social Media Feeds )
This submission is an attempt to solve these business problems by leveraging the power of Cognizant’s Pioneering SMAC (Social - Mobile - Analytics - Cloud) Framework
Our Solution
Challenges
© 2014 Cognizant 4
Introduction“SpendWise Genie”, is a multipurpose Mobile App that gives consumers rich insights into benchmarking their spend behavior with people of similar demographic profile( by location and merchant category). The App also uses the power of real time predictive algorithms for relevant offer presentment and empowered decision making
BBVA App For Customers
Value Proposition
SpendWiseGenie
Spend Comparison with
Peers across segments Responsible
Spending (for Consumers*)
+Increased Sales
(for Banks & Merchants)Best Offers
Prediction supported with real-time data ( Maps/ratings)
SpendWise. Be wise
People Like Me
Offers For Me
Leverages the power of BBVA API’s
Non BBVA customers can also benchmark their spend
behaviour through this App
© 2014 Cognizant 5
User Journey
SpendWise Genie is a Smart “Spend Benchmarking & Real Time Offer Presentment” App that leverages BBVA APIs & advanced predictive algorithms
View
• Minimal input details (Age group, Gender etc.) for Non BBVA customers
• Interactive charts & visualizations to highlight spend patterns across category-location- time continuum
1
Plan
• Empowered spend planning across categories and peer groups
3
Compare
2
• Compare own spend pattern with peer segment and other BBVA Customers to understand deviations or identify high spend categories
Explore
4
• Receive relevant offers and reduce spend (Prediction based on consumer preferences, merchant location, discount offered etc. )
© 2014 Cognizant 6
Key Insights
“SpendWise Genie” is designed to answer the following:-
What is your spend pattern across different
spend categories & time-periods?
How does your spend compare to your peer
segment?
How can you plan & optimize your spend across categories ?
What are the most relevant offers which
you can utilize ?
© 2014 Cognizant 7
“People like me” AppWith “People Like Me” learn how consumers spend in your same category
People Like Me
Rich interactive charts represent the spend pattern across time cohorts
The consumer also picks spend category and subcategoryThe user provides his age, gender & location information and also the time period for which he wants to view spend patterns
For App demo refer to
pps file
© 2014 Cognizant 8
The “Offers for me” AppWith “Offers For Me”, review relevant and timely offers with supporting information ( maps, reviews and Ratings)
Offers for me
The user provides his age, gender & location information
The consumer specifies anticipated spend range or can be presented offers for the selected spend category
The selected merchant location is displayed on an interactive map where consumer has the option to get driving directions from current location (GPS)
The consumer can also view ratings and reviews of the selected merchant, available with ‘Google Places’
The top merchant offers pertinent to the spend category and anticipated spend range are presented to the consumer as a listing
For App demo refer to
pps file
© 2014 Cognizant 9
Key Business Benefits delivered
Promotes Responsible Spending - Know whether you are over spending / or can afford to spend more in certain categories
People Like Me
Offers For Me
Save on spend by redeeming offers
Drive / Reach new customers (by motivating those who underspent in specific categories)
Drivel Sales through increased Footfall and attract new customers
Design Future Offers/ Campaigns based on the performance / take-up of offers
Deliver the power of data to customers and increase loyalty by being a source of personal finance planning information
Drive Spend – thus increasing benefits for the bank and its merchants
Increased Customer Loyalty through an effective Offer Presentment Program
Consumer Merchant Bank
© 2014 Cognizant 10
Solution Architecture Business & Data Flow
BBVA Data
AnalyticsBBVA Host
BBVA API’s Recommendation rules to push relevant offers based on offer
score & customer segment mapping
Easy Visualization of Spend
Patterns
Existing Consumer
Transaction
Aggregated BBVA Consumer Data | Individual Consumer Data
Review Feed
Analyzer
Regular feed of Social Posts Google
Places APIs
High prediction accuracy for
spend category & range
Merchant locations with favorable ratings & reviews
SpendWise. Be wise
Advanced Predictive Model
Cloud Hosted Model
© 2014 Cognizant
Services
11
Technical Architecture Application Design / Details
Mobile Android Application
Ionic js
Angular JS
Cordova / Phone gap
Offers Service hosted in
Google app engine
Prediction Service via Advanced predictive
models
BBVA Data API
Utilities
Auditing
Logging
Caching
Exception
Handling
Con
necti
vit
y
Locator Service via
Google Maps API
Social Sentiments via Google Places
API
Solution Building
BlockDescription
Mobile Android application
UI was developed using ionic and angular js. High charts was used as charting framework. Mobile app
was packaged using Cordova.
Offers ServiceThis service was built to filter and display the offers
specific to user interest. This service is hosted in Google app engine.
Prediction Service
This service was built using Advanced Predictive Models.
Locator ServiceDetails of the merchant providing the offers are
located on Maps and direction from user’s location to merchant location is provided.
Social Sentiments
Reviews and ratings of the merchant are provided using Google Places API
Utilities Common components to address non functional related common concerns across layers
High charts
© 2014 Cognizant
Cloud Based Predictive Model Design
Modeling
Business Understanding
DataUnderstanding
DataPreparation
Evaluation
Deployment
Objective of the model is to predict user spend category and spend amount based on:- The historic spend behavior of user segment (age group & gender);- Location (zip code) based spend patterns as well as seasonality trends – based on day-of-the-week and time-of-the-
dayThe base data* used for model training is based on the BBVA transaction data for:- Top 10 spending Zip Codes using BBVA APIs: Cards Cube and Consumption Pattern- Data filtered for Top 5 Merchant Categories with the highest spend in these 10 zips
Input data* for predictive model training is prepared by combining data from above 2 BBVA APIs to calculate the most probable spend on merchant category for a user segment at a weekday, at a particular hour: - By calculating merchant spend probability at an hour of a weekday using weighted contribution of spend at that hour
and spend by that user segment- For the sake of simplicity, spend behavior on a particular weekday (e.g a Tuesday) across months is assumed to be
similar
Using Advanced Predictive Models for a given Zip Code, User Segment , Day, Hour of Day:- Model 1: Prediction of Merchant Category with the Highest Probability of next User spend; Model Type: Classification Model; Model Accuracy: 89%- Model 2: Prediction of Spend Range for a user for selected/predicted Merchant Category; Model Type: Regression
Model; Refer MSE (Mean Square Error Values in Appendix)
Model Output i.e. the predicted category and the predicted spend is used to pull offers from offers database filtered for :- Selected Zip Code, Merchant Category, Predicted Spend (Range)- Pull rating/reviews from Google Places for displayed filtered offers
Detailed Evaluation of Models explained in next 2 slides
12
CRISP – DM Methodology Followed
* - All local copies of data made to test model accuracy have been purged
© 2014 Cognizant
Offer Prediction Model
13
Prediction Models
Prediction of the Merchant Category with the Highest Probability of next User spend in a given:- Zip Code- User Segment (Age Group, Gender)- Week Day- Hour of Day
1
Spend Prediction
Category Prediction
Prediction of the Spend Range for a user for selected/predicted Merchant Category in a given:- Zip Code- User Segment (Age Group, Gender)- Week Day- Hour of Day
2
Model
Model
Model Limitations:Since the current model uses the costumer transaction data as input data from BBVA APIs , which is available only at a segment level (age-group, segment) and not at a customer ID level, the models output is only valid for segment/cohort level predictions, assuming all the customers within that cohort behave in a similar spending manner.
Cloud Based Predictive Model Design
© 2014 Cognizant
How it works…
Prediction of Merchant Category with the Highest Probability of subsequent user spend in a given:- Zip Code - User Segment (Age Group, Gender)- Week Day - Hour of the day
Assumptions:- Training* and prediction limited to Top 5 Spend Merchant Categories (except mx_others):
{(mx_basrsandrestaraunts, mx_food, mx_fashion, mx_auto, mx_hyper (mall)}- High Accuracy prediction for Top 10 spending zips for current model
Model Type: Classification Model; Model Accuracy: 89%Test Cases:
Zip Code
Age Group GenderWeek Day
HourActual Category with
Maximum SpendPredicted Category
"11000" "19-25" "Female" "Fri" 5 " mx_barsandrestaurants1" " mx_barsandrestaurants1"
"11000" "26-35" "Unknown" "Mon" 19 " mx_food1" " mx_barsandrestaurants1"
"11320" ">=66" "Male" "Thu" 1 " mx_food1" " mx_food1"
"11510" ">=66" "Male" "Mon" 17 " mx_travel1" " mx_food1"
"11520" "Unknown" "Female" "Thu" 19 " mx_fashion1" " mx_barsandrestaurants1"
"14300" "19-25" "Female" "Fri" 16 " mx_hyper1" " mx_hyper1"
"11510" "Unknown" "Female" "Mon" 10 " mx_travel1" " mx_travel1"
"11000" "56-65" "Female" "Tue" 6 " mx_beauty1" " mx_barsandrestaurants1"
"11000" "Unknown" "Female" "Fri" 1 " mx_beauty1" " mx_beauty1"
"11590" "<=18" "Male" "Sat" 2 " mx_auto1" " mx_auto1"
1Model
14
Evaluation of the Data Mining Model for Spend Category Prediction
* - All local copies of data made to test model accuracy have been purged
© 2014 Cognizant
Prediction of Spend Range for a user for selected/predicted Merchant Category for:- Zip Code - User Segment (Age Group, Gender)- Week Day - Hour of Day
2Model
Assumptions:- Training* and prediction limited to Top 5 Spend Merchant Categories (except mx_others):
{(mx_basrsandrestaraunts, mx_food, mx_fashion, mx_auto, mx_hyper (mall)}- High Accuracy prediction for Top 10 spending zips for current model
Model Count: Separate Model for each of the Top 5 Merchant Categories
Model Type: Regression Model; Test Cases: E.g. For merchant Category – “mx_auto”
Zip Code
Age Group Gender Week Day Hour Actual Spend Predicted Spend
"11590" "19-25" "Female" "Fri" 7 147.33 483.6822
"11320" "36-45" "Male" "Mon" 17 1114.3 1611.875
"11520" "46-55" "Male" "Sat" 12 2723.5 1415.509
"64000" "56-65" "Male" "Wed" 18 168.15 169.7097
"11590" "46-55" "Male" "Wed" 9 1000 1294.2
15
How it works….Evaluation of the Data Mining Model for Spend Range Prediction
* - All local copies of data made to test model accuracy have been purged
© 2014 Cognizant 16
Recommended Roadmap…. All zips from BBVA consumers and across Mexico can be incorporated
All the merchant categories can be included in analysis
Real time access to offers database
Real time Social media sentiment analysis to push relevant places / offers to specific segments
Customer level transaction data , if made available, more specific & targeted offers can be built using more sophisticated algorithms
Customers can like/dislike offers to generate valuable insights for future offer design
Incorporation of external data factors (e.g. weather data) to suggest suitable offers / merchants
Integration of payments/offer redemptions through Dwolla, helping merchants track & plan offers
iOS version of the app would be launched
A Spanish version of the app in agenda for future release
© 2014 Cognizant 17
Solution evaluation parameters
Originality
The app is uniquely positioned to address
customer need of spend optimization
through relevant offer presentment based on peer-group past
spend behavior, location, day of the week & time of the
day
Visual Appeal
The app enables the customer to view his peer segment spend
patterns across spend categories at
different drill down levels through rich interactive charts. Also, once a user selects an offer -
relevant merchant details, ratings, reviews, social
sentiment and its location , directions on an interactive
map are presented
Usefulness
1. Spend Tracking & Reporting
2. Peer Benchmarking
3. Spend Prediction & Offer Presentment
4. Spend optimization
Usability across devices
The app uses consumer’s current location along with
other segment parameters to report & recommend offers
based on cloud hosted Advanced Prediction API rule engine. Mexican
mobile market has a prevalence of Android OS (>70% share *),
the app can be currently used across
Android mobiles & tablets
External Data
1. Google Places API (ratings & reviews)
2. Google Maps API 3. Offers database
as a proxy for bank offers / Groupon data
* Source - http://www.statista.com/statistics/245193/market-share-of-mobile-operating-systems-for-smartphone-sales-in-mexico/
© 2014 Cognizant 18
Appendix
© 2014 Cognizant 19
Mocked Up Offers Database
The offers database consists of following variables: Actual Merchant Details in Mexico
• Merchant Name• Merchant Address• Merchant Zip Code• Merchant Latitude & Longitude• Merchant Sample Image Link• Merchant Category
Mocked-up Offers
• Offer Details Spend Amount in Mex$ Discount% Savings in Mex$
Sample offers data
© 2014 Cognizant
Predictive Model Evaluation (1/2)
Model Description # of Instances Model Type Classification Accuracy
Model 1: Category Prediction 21,161 ClassificationIn Sample Validation: 89%
Out Sample Validation: 64%
Model Description # of Instances Model TypeSquare Root of Mean Squared
Error (Develpoment Sample)
Model 2: Spend Prediction
(mx_auto)72,737 Regression 1,824
Model 2: Spend Prediction
(mx_barsandrestaraunts)361,616 Regression 1,470
Model 2: Spend Prediction
(mx_hyper)133,114 Regression 5,036
Model 2: Spend Prediction
(mx_food)253,337 Regression 1,017
Model 2: Spend Prediction
(mx_fashion)79,648 Regression 2,557
20
© 2014 Cognizant
Predicted Category mx_auto mx_barsandrestaurants mx_hyper mx_food mx_fashion
mx_auto 100% 0% 0% 0% 0%
mx_barsandrestaurants 0% 93% 1% 6% 0%
mx_hyper 0% 0% 84% 16% 0%
mx_food 0% 8% 2% 89% 0%
mx_fashion 0% 98% 0% 0% 2%
Prediction Training Confusion Matrix: In Sample ~ 89%1Model
High Prediction Accuracy for 4 merchant categories
Poor Prediction Accuracy for 1 merchant category
21
Predictive Model Evaluation (2/2)
Predicted Category
mx_auto
mx_barsandrestaurants
mx_hyper
mx_food
mx_fashion
Low
Low
Prediction Accuracy: Out Sample Validation ~64%
Accuracy Level
High
High
High
1Model
Based on 50 Out Sample Validation Test Cases
© 2014 Cognizant
Visit link for app download :- https://bbvaopen4u.cognizant.com/SpendWiseGenie
Visit link for app demo :- https://bbvaopen4u.cognizant.com/SpendWiseGenie/SpendWiseGenieDemo.ppsx22
Thank you
top related