General Insurance Company
Fraudulent behavior detection @ point of insurance policy login by
- Identification of groups indulging in fraud (internal & external)
- Identification of adverse selection of lives & deceitful claims
The biggest challenge faced by insurance companies these days are fraudulent transactions/ policies. Customers employ various ways to make fraudulent claims in terms of over insurance, staged collisions for claims, exaggerated claims, etc sometimes in tandem with the agents.
Most of these cases are identified only after the claims are issued leading to companies incurring loss due to early claims, but also higher premium prices for the low risk groups. The aim is to identify potential claimants at the point of signups even before the policy is issued.
Scalend Insights Appliance uses Hadoop as a cost effective platform for handling ever growing volumes & variety of data. Strong encryption techniques are incorporated to ensure adept masking of PII data liaising with the PCI compliance.
Data Utilized –Real time Clickstream data from web & mobile apps, individual user data, social media data from the users, and internal data such as agent information.
The data is brought into the system through the ready to use connectors, cleansed and modeled based on industry standard algorithms to gain insights.
Algorithms Used – Scalend Insights Appliance uses the K-means clustering algorithm to identify matching patterns within the customer & agent data. The customers & agents are categorized into high, medium and low risk profiles.
This, along with the risk scores for agent & customer identified the prospective customers / agents causing false claims.
Leading Stock Broking Firm
Real time social profiling of customers
With large number of players coming into investment broking arena, customers always look for the firms offering the best solution & guidance for their needs. The companies are trying to leverage the technological advantages for getting real time insights about its customers to provide personalized services.
The requirement was to get near real time profiling for new and existing customers based on attributes such as demography, income bracket, age and available social data (Facebook, Twitter, Instagram, Linkedin) & app usage for personalized recommendations/suggestions on stocks.
Scalend Insights Appliance uses its crawl feature as a part of the text mining to fetch data from social media user profiles. Other than this, the click stream data is also utilized for real time insights. The data is stored in the Hadoop data lake with the required encryption & masking features
Data Used – Form data (details entered by the user in the website forms), social media data & real time clickstream data from web and mobile apps to analyze and give accurate insights into the customer choices.
Sentiment Analysis – Based on the basic details entered in the website forms, the social media data is
fetched at near real time. The user sentiments such as page likes, dislikes and comments are captured and synthesized from a wide range of social media platforms such as twitter, facebook, instagram, etc.
This helps identify the key sentiments, customer choices & feedback.
The social media data is then joined with the form data and click stream information such as the users’ page of interest from websites or mobile apps. The data is cleansed & modeled using industry standard statistical algorithms for user grouping
Algorithms – The users are classified into different groups based on various factors such as sentiment score, clickstream data, etc. K-means clustering algorithm is used to classify customers into various clusters based on which new and existing users can be given personalized suggestions on the investments almost at real time.
Online Cashflow Management Company
Customer Retention / Churn Analytics
Improve customer conversion rates & retention and maximize the customer life time value. Understand the customers to run targeted campaigns, thus reducing customer churn & acquisition costs.
Along with the regular page traffic statistics, the customer insights platform tracks every event from the time the user enters the website (anonymous) and maps it with the events performed after the user signs up (converted user).
The insights also provided the idea on the most successful campaigns and the user conversion from the respective campaigns (social media, personalized emails, etc).
User Churn is calculated from the tracked events to help evaluate targeted campaigns to bring inactive customers back in its fold by understanding the reason for their drop off.
The customized & detailed dashboards for the user information were provided in less than a day of implementation.
Leading Payment Merchant
With the rise in e-commerce & more emphasis given on digital transactions by the regulatory authorities, most of the transactions online or in brick and mortar outlets are happening digitally. Merchants need to gain more insights into the customer needs in order to provide better service to the users.
Scalend Customer Journey Insights is a simple SDK that can be installed in the merchant system to get insights from its users. The SDK tracks the user journey almost immediately after the insertion. The solution also brings in social media data through its crawl features to provide better understanding of the competitors and the customer choices. All the data is combined & brought together into a single data lake for further analysis.
Data Used –Merchant web & mobile app data, Social media data, historical data
Social Media Analysis – is performed to understand the general landscape of the merchants & their competitors and get the sentiment of the customers wrt each of their services to imap the online reputation. >
Customer Journey Insights – Helps get real time insights from the web, mobile & other appliance such as POS for faster and smarter decision making on customer engagement and retention.
With all the data in a single store and ready to use analytics platform, it is much easier to review the transactions from various touch points & predict the revenue in the coming holiday, seasons, weather and various other factors.
Card Fraud Analysis
Industry has been seeing a rise in the fraudulent transactions using debit/ credit cards owing to the rise in the digital transactions. The credit & debit card issuers need to act on an immediate basis to avoid potential loss, thus preventing negative impact on the customer confidence & reputation of the brand.
Organizations holding on to card & PII information of customers have varied levels of data security features. Hence, it becomes necessary to ensure that every access point is secure.
Scalend Insights Appliance's ready to use and scalable data lake connects & brings in data from various touch points into a single data store at near real time to get faster insights.
Customer Journey Insights was also implemented in order to get a user level understanding and detect any anomalies in their regular behavior, such as sudden high value spending & other suspicious activities.
Algorithms – User grouping based on various factors such as their spending choices, location, etc is done using K-means clustering. Regression techniques are run to track & get predictive insights and early warning signals of fraud.