Analytics for Matrimonial
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Analytics for
Matrimonial
AAUM Confidential
This document contains information and data that AAUM considers confidential. Any disclosure of Confidential Information to, or use of it by any
other party, will be damaging to AAUM. Ownership of all Confidential Information, no matter in what media it resides, remains with AAUM.
Corporate profile
Founded by IIT Madras alumnus having extensive global business experience with Fortune 100
companies in United States and India having three lines of business
Analytics Competitive Intelligence Livelihood
• Appropriate statistical models • Actionable insights to clients for •Services ranging from promotion of
through which clients can measure their business excellence livelihoods, implementation services,
and grow their business. livelihood & feasibility studies.
Key Focus Areas in Advanced analytics and Predictive analytics
Product – geniSIGHTS (Analytics/BI), Ordo-ab-Chao (Social Media)
More than 25 consulting assignments for Businesses & Govt orgs
Partnership – Actuate, IIT Madras, TIE and 3 strategic partnerships Aaum’s office, IIT Madras Research Park
Dedicated corporate office at IIT Madras Research park since 2009
Prof Prakash Sai Puneet Gupta Padma Shri Dr. Ashok Jhunjhunwala
Dr. Prakash Sai is professor at the Department Puneet spearheads the IFMR Mezzanine Dr. Ashok Jhunjhunwala is Professor at the
of Management Studies, Indian Institute of Finance (Mezz Co.), is strengthening the Department of Electrical Engineering, Indian
Technology Madras. He has wealth of delivery of financial services to rural households Institute of Technology Madras India. He holds a
international consulting experience in Strategy and urban poor by making investments in local B.Tech degree from IIT, Kanpur, and M.S. and
Formulation financial institutions. Ph.D degrees from the University of Maine, USA.
-2-
Competencies in
Advanced analytics Competitive intelligence Livelihood
Competitive
assessment
Build appropriate statistical models through Provide actionable insights to clients for Perform livelihood services ranging from
which clients can measure and grow their their business excellence. promotion of livelihoods, implementation
business. services, livelihood and feasibility studies.
Expertise in Expertise in Expertise in
• Digital Media •Business Entry •Government organizations
• Finance/Insurance •Business Expansion •Non Government
• Retail •Market research organizations
• Entertainment •Corporate with livelihood
• Human Capital focus
• Government organizations •Research
• Research & training
-3-
Corporate profile
Founded by IIT Madras alumnus having extensive global business experience with Fortune 100
companies in United States and India having three lines of business
Analytics Competitive Intelligence Livelihood
• Appropriate statistical models • Actionable insights to clients for •Services ranging from promotion of
through which clients can measure their business excellence livelihoods, implementation services,
and grow their business. livelihood & feasibility studies.
Competitive Advantage –Advanced analytics, Predictive analytics
Executed more than 25 consulting assignments for Businesses and Govt organizations
Team of 20 members
Dedicated corporate office at IIT Madras Research park since 2009
Prof Prakash Sai Puneet Gupta Padma Shri Dr. Ashok Jhunjhunwala
Dr. Prakash Sai is professor at the Department Puneet spearheads the IFMR Mezzanine Dr. Ashok Jhunjhunwala is Professor at the
of Management Studies, Indian Institute of Finance (Mezz Co.), is strengthening the Department of Electrical Engineering, Indian
Technology Madras. He has wealth of delivery of financial services to rural households Institute of Technology Madras India. He holds a
international consulting experience in Strategy and urban poor by making investments in local B.Tech degree from IIT, Kanpur, and M.S. and
Formulation financial institutions. Ph.D degrees from the University of Maine, USA.
-4-
Competencies in
Advanced analytics Competitive intelligence Livelihood
Competitive
assessment
Build appropriate statistical models through Provide actionable insights to clients for Perform livelihood services ranging from
which clients can measure and grow their their business excellence. promotion of livelihoods, implementation
business. services, livelihood and feasibility studies.
Expertise in Expertise in Expertise in
• Digital Media • Business Entry • Government organizations
• Finance/Insurance • Business Expansion • Non Government organizations
• Travel & Logistics • Market research • Corporate with livelihood focus
• Retail
• Research
• Entertainment
• Human Capital
• Government organizations
• Research & training
-5-
What Aaum can offer
Performance Recommendation
analysis engine
Insightful Advanced
dashboard Analytics
Experience the magic by clicking our Genie
-6-
Performance Recommendation
analysis engine
Performance analysis at various levels are vital to the success Insightful Advanced
dashboard Analytics
Churn
GENERATE
Registrants
Partially completing Free Expresses
the free profile registrants interest for Paid registrants
paid
COLLECT Feedback
Churn
Right KPIs tied to right customer profiles at right time
-7-
Performance Recommendati
Recommendations based on the learnings from the analysis on engine
performance analysis Insightful Advanced
dashboard Analytics
•What approach for what candidate profile
• Conversions (Likelihood to make the payment)
• Targeting right prospects
• Page access
• Time spent on tasks
• Submission of horoscopes, photos, references, etc
• Time spent on profiles
• Responses, acceptance, rejects, interests
•Advise/empower the “personal relationship manager” to adopt
appropriate preventive/corrective actions
• Kiosk/Center level recommendations
• What profiles to approach for quick conversions
-8-
Performance Recommendatio
analysis n engine
Insightful dashboard to improve the business Insightful Advanced
dashboard Analytics
• Monitoring the candidate specific score levels
•Make actions from it
• Categories
• forming elite communities IIT, IIM, etc
• IT software developer from top B schools with eng
Background
• Caste/Sub caste/Communities
• Category performance vs overall performance.
• Overall performance to median (Bellcurve)
• Trend analysis
• and many …
Strategic partnership with
• state of the art reporting solutions
-9-
Performance Recommendation
analysis engine
Advanced analytics to illuminate insights from the data
Insightful Advanced
dashboard Analytics
• Predicting prospect’s willingness to continue
• Churn
•Acquisition
• Cohort analysis to know the effectiveness of the engagement
• Planning and Scheduling promotions/events
• “Generic online Sunday meetings” vs Café Coffee Day meetings for a few”
• Building models that watch for thresholds
•Early alerts i.e. keeping track of attrition
• Remove fake profiles
• Based on advanced algorithms
- 10 -Free state activity - approach
Free period is computed, which is the difference in days between the next payment date and previous expiry date.
Free period is bucketed into the levels,
• 0-15 days(level 1), 15-30 days(level 2), 30-60 days(level 3), 60-365 days(level 4), >365 days(level 5).
Activities are classified into 6 broader categories,
Login
Basic profile modification
• Assured contact, Birthday offer, Family info modification, Hobby modification, Modification screening, Profile modification
Advance profile modification
• Add photos, Add reference, Horoscope generation, Ignore, Mobile alert registration, Profile delete
Match view
Match initiation
• Contact history, Filter, Match watch, Photo match watch
Match communication
• Block, Horoscope request receive, Horoscope request sent, Online payment failure, Personal message receive, personal
message sent,, Phone view request receive, Phone number request sent, Interest receive, Reference request receive, Reference
request sent, Sent interest, Voice message receive, Voice message sent, Phone number request receive
- 11 -Free period – looking into the distribution(days) & Customer segments
Free period is the difference in days of previous expiry Free period of the customers been divided into
- 0-15 days (level1), 15-30 days (level2), 30-60 days
date and next payment start date. (level3),60-365 days(level4), >365 days (level5)
The above plot gives the distribution of the free period, the
free period is ranging from a min of 2 days to max of 1563 No free period in the level4, i.e, 60-365 days.
days, with an average free period of around 130 days.
The plot gives us the proportion of the customers coming
The second plot gives a better view of the distribution. under each level of the free period.
Around 80% of the customers became paid members
with in 2 months of free period.
- 12 -Activity in the free period levels – proportion of the customers
Login and Match view are umbrella activites, hence not considered while dwelling deep into the activities.
- 13 -Activities, sub activities in the free period
Add photo accounts to 60% of the advance
Profile modification accounts to 57%
modification
of the Basic modification
Around 88% of the match initiation type is The activities – personal msg receive, receive
due to the photo match activity interest and sent interest account to almost 71.5%
of the match communication activity category
The above plot is the broad activity type and underlying activity view.
- 14 -Free period levels – various activities
We have come up with a metric – Expected activity per day, to measure the activity of each customer
coming under various free time levels.
The metric is the ratio of the no.of times the activity happened to the free time days per each customers
each activity.
This metric is calculated at the broader activity type and at even each of the activities.
The following slides will help in understanding the expected underlying activity distribution over the
different free time levels.
Activity period in days, calculated as the difference from next payment date and activity date.
Further, the activity period is divided into three phases, namely “Initial phase”, “Mid phase”, “Near
payment phase”
Initial phase – first one-third of the activity period
Medium phase – second one-third of the activity period
Near payment phase – last one-third of the activity period
Further dwelled deep by looking into 5 parts of the activity period categorised as “Phase 1”, “Phase 2”,
“Phase 3”, “Phase 4”, “Phase5”
- 15 -Activity phases in the free period (0-15 days) – various activities
The free period is categorised as “initial
phase”, “mid phase”, “near payment
phase”,
There is sudden dip in the average
expected underlying activity from the
second phase to the third phase which is
the near payment phase and this tells us
that the underlying activity has been more
performed more in the start and mid phrase
of the free period.
- 16 -Activity phases in the free period (15-30 days) – various activities
The free period is categorised as “initial
phase”, “mid phase”, “near payment
phase”, which are the 1/3rd, 2/3rd & 1 of
the free period.
- 17 -Activity phases in the free period (30-60 days) – various activities
The free period is categorised as “initial
phase”, “mid phase”, “near payment
phase”, which are the 1/3rd, 2/3rd & 1 of
the free period.
- 18 -Activity phases in the free period (>365 days) – various activities
The free period is categorised as “initial
phase”, “mid phase”, “near payment
phase”, which are the 1/3rd, 2/3rd & 1 of
the free period.
- 19 -5 Phased activity period, free period (0-15 days) – various activities
The free period is categorised as “phase1”,
phase2”, “phase3”, “phase4”, “phase5”,
which are the 1/5th, 2/5th 3/5th, 4/5th,1 of
the free period.
- 20 -5 Phased activity period, free period (15-30 days) – various activities
The free period is categorised as “phase1”,
phase2”, “phase3”, “phase4”, “phase5”,
which are the 1/5th, 2/5th 3/5th, 4/5th,1 of
the free period.
- 21 -5 Phased activity period, free period (30-60 days) – various activities
The free period is categorised as “phase1”,
phase2”, “phase3”, “phase4”, “phase5”,
which are the 1/5th, 2/5th 3/5th, 4/5th,1 of
the free period.
- 22 -5 Phased activity period, free period (>365 days) – various activities
The free period is categorised as “phase1”,
phase2”, “phase3”, “phase4”, “phase5”,
which are the 1/5th, 2/5th 3/5th, 4/5th,1 of
the free period.
- 23 -Basic modification, underlying activities – (0-15 days)
Further looking into the underlying activites of each broader activity classification. The above plot is the distribution of the various expected
activity distribution over the three phases of the free period of the activity type “Basic profile modification”.
- 24 -Advance modification, underlying activities – (0-15 days)
The above plot is the distribution of the expected activities distribution over the three phases of the free period of the activity type “Advance
profile modification”. Here we can see the average add photos, add referance, and ignore are high in the mid phase.
- 25 -Match initiation, underlying activities – (0-15 days)
The above plot is the distribution of the expected activities distribution over the three phases of the free period of the activity type “Match
initiation”. Here we can see the average of expected contact history and match watch are stable in the first two phases and dropping in the
third phase. Photo match watch average expected activity is dropping down over the phases.
- 26 -Match communication, underlying activities – (0-15 days)
Most of the activities remained stable over the three phases except for the activity interest receive
and sent interest where the average dropped in the near payment phase.
- 27 -Basic profile modification, underlying activities – (15-30 days)
Basic profile modification, 15-30 days, the distribution of the underlying expected averages.
- 28 -Advance profile modification, underlying activities – (15-30 days)
Advance profile modification, 15-30 days, the distribution of the underlying expected averages.
- 29 -Match initiation, underlying activities – (15-30 days)
Match initiation, 15-30 days, the distribution of the underlying expected averages.
- 30 -Match communication, underlying activities – (15-30 days)
Match communication, 15-30 days, the distribution of the underlying expected averages.
- 31 -Basic profile modification, underlying activities – (30-60 days)
Basic profile modification, 30-60 days, the distribution of the underlying expected averages.
- 32 -Advance profile modification, underlying activities – (30-60 days)
Advance profile modification, 30-60 days, the distribution of the underlying expected averages.
- 33 -Match initiation, underlying activities – (30-60 days)
Match initiation, 30-60 days, the distribution of the underlying expected averages.
- 34 -Match communication, underlying activities – (30-60 days)
Match communication, 30-60
days, the distribution of the
underlying expected averages.
- 35 -Basic profile modification, underlying activities – (>365 days)
Basic profile modification, >365 days, the distribution of the underlying expected averages.
- 36 -Advance profile modification, underlying activities – (>365 days)
Advance profile modification, >365 days, the distribution of the underlying expected averages.
- 37 -Match initiation, underlying activities – (>365 days)
Match initiation, >365 days, the distribution of the underlying expected averages.
- 38 -Match communication, underlying activities – (>365 days)
Near payment phase Near payment phase Near payment phase Near payment phase
Near payment phase Near payment phase Near payment phase Near payment phase
Near payment phase Near payment phase Near payment phase
Near payment phase Near payment phase Near payment phase
Match communication, >365 days, the distribution of the underlying expected averages.
- 39 -Analytical Advantage Using Mathematical modeling
Case studiesCase Study : Campaign and Publisher scoring for a major US based digital media &
content platform firm
Project Context: Our client, a US based media and content platform firm for engaging the audience with the most money and
influence online with more than 1,000 publishers as partners. The client approached Aaum to reap analytical insights from
their humongous transaction data
AAUM’s Contribution:
• Devised and tracked the Key proportions and Key Performance across the sites and publishers
• Achieved “performance scores” to qualify the site/campaign performance characteristics (CTR, Margin, CPM and Delivery rate) and got validated.
• Extended the scoring technique to derive segment specific indices and compared the peer campaigns
• Root cause analysis: Inspection of site performance attributes of the critical campaigns
Sample Deliverables
… scoring is extended further to derive segment
… Key proportions and Key Performance … qualifying the performance by scoring specific indices and compared with their peers
Root cause analysis of the critical campaigns
… acid test on our scores
- 41 -Case Study : Analytical support for a major television network
Project Context:
Our client, a national television network approached us to derive analytical insights by slicing and dicing
multiple scenarios to help the top management to arrive at actionable results.
AAUM’s Contribution:
• Understanding of target audience, frequent respondents and provision to reward the most loyal respondent (Integrated over a period of time)
• Respondent behavioral characteristics - by age group, gender, voice/non voice
• Media campaign/promotions efficacy - What time is better, Which channel is better?
Sample Deliverables
… extracted insights by … chennai’s female population ….analysis strengthened the
dimensional analysis are older than other vital media channel to reward the
circles loyal customers.
- 42 -Case Study: Developed KPI’s and Predictive Analytics for a HR Consulting firm
Project Context:
Our client, a HR consulting firm, approached AAUM to develop comprehensive metrics which could be
standardized across its clients
AAUM’s Contribution:
• Developed standard metrics that could be rolled out across to the clients of the consulting firm
• Performed interaction analysis for ‘key metrics’ to derive a holistic understanding
• Developed predictive models for some very useful parameters
Sample Deliverables
Standardized metrics were Interaction analysis of key Predictive models were built
defined and developed metrics was performed and using Random Forests for key
insights were derived parameters
- 43 -Case Study: Analytical support for a major ticket booking company
Project Context:
Our client, a major ticket booking company approached AAUM to provide comprehensive analytical insights to
help the top management to arrive at actionable results.
AAUM’s Contribution:
• Developed RFM metrics to help the management to effectively segment customers and reward them.
• Performed cohort analysis to qualify the churning, effective engagement with agents.
• Qualified traveler’s behavioral patterns on routes, regularity, age, gender, per ticket value, lead days, etc.
• Developed dynamic pricing strategy based on the unfilled inventory levels.
Sample Deliverables
Margin analysis … Cohort analysis to qualify churn & engagement disengagement levels …
RFM … acquisition …
- 44 -Case Study: Analytical support for the largest integrated travel and travel related
financial services company
Project Context:
Our client approached AAUM to offer effectively qualify their business operations with insights from data analysis.
AAUM’s Contribution:
• Devised methodology to effectively qualify the payment cycle, ageing metrics, outstanding from customers.
• Performed credit limit analysis, qualified credit consumption patterns and suggested efficient methodologies.
• Performed fare analysis to optimize the travel offerings to customers.
Sample Deliverables
lead analysis on ticket fare of credit consumption and breaching promptness in payments …
airline travel … patterns …
- 45 -Case Study: Lodging analytics for leading global lodging solutions company
Project Context:
AAUM offered analytical support to a leading global lodging solutions company to leverage their business.
AAUM’s Contribution:
• Slicing and dicing of data to understand customer’s hotel booking preferences qualified on savings, effective
rate and net paid.
Sample Deliverables
effective rates over customer preference, location,
Trend analysis to understand the effective rate pattern in
position, etc …
a year …
- 46 -Case Study: Business intelligence for a leading financial services company
Project Context:
Developing Business intelligence and analytical capability for a leading financial service company
AAUM’s Contribution:
• Analyzed current BI maturity with comprehensive data analysis of the customer datum collected by our client
• Evaluated and suggested comprehensive plan for BI maturity
• Identified as strategic partner to perform predictive analytics in multiple phase spanning two years
Sample Deliverables
- 47 -Case Study: Measuring the movie performance by mining social media sentiments
Project Context:
Research project funded by the company to devise a performance dashboard on various attributes for a movie to predict the
success at the Box office.
AAUM’s Contribution:
• CaseObjectives
Team AAUM extracted and analyzed the information from the web sources like YouTube and twitter to achieve the specified Business Study :
• Developed dashboards for various attributes and analyzed the trends in the sentiment scores by using clustering, classification techniques.
Sample Deliverables
-
4 -7-
- 48 -Case Study: Market Basket Analysis and Loyalty program
Project Context:
Our client approached us to develop a framework to analyze the transaction data of their customers and tie these insights with loyalty
card program and reward their customers by their preference and loyalty factor.
AAUM’s Contribution :
• Mined association rules to provide strategic insights for cross-sell/up-sell opportunities
• Framed out loyalty management model for the client.
Sample Deliverables
- 49 -Case Study: Crime Analytics for Tamil Nadu State Crime Records Bureau
Project Context:
Our client, police Department archived the crime data for the past ten years and approached AAUM to deliver
insights by performing appropriate analytical techniques.
AAUM’s Contribution:
• Discovered hotspots of crimes by cluster analysis of the historic data.
• Trend analysis of different crimes
• Mining insights from the police records by text mining
• Derived KPIs for crime department to monitor the crimes.
Sample Deliverables
Mining insights from the comments recorded
Hot Spots found out for crimes in Chennai city.
Trend movement for each type of Crime
- 50 -Case Study: Predictive model to grade the performance of DCCBs
Project Context:
Developing Business intelligence and analytical capability for a leading rural bank
AAUM’s Contribution:
• Analyzed current BI maturity with comprehensive data analysis of the customer datum collected by our client
• Evaluated and suggested comprehensive plan for BI maturity
• Identified as strategic partner to perform predictive analytics in multiple phase spanning two years
Sample Deliverables
- 51 -Case Study : Analytics on data collected by a rural research firm
Project Context:
Our client has collected valuable data through their Kiosks and wanted to derive insights which can identify new potential business
for rural India
AAUM’s Contribution:
• AAUM profiled the villages and identified clusters
• Thorough analysis done for seven villages in three districts based on income slabs adopted from NCAER & Mckinsey report
• Identified new businesses which would flourish once the villagers migrate to higher income slabs
Sample Deliverables
As the villagers start earning More than 50% of
….and they prefer private villagers in Allivilagam
more they will spend on
hospitals. earn less than Rs.
communication and health
care …. 22,500 per annum
Never visited doctor
0.32% 0.00%
0.43%
6.35% 0.84%
0.42% 0.33%
0.00%
0.67% 0.00%
1.35%
0.00%
1.35% 0.00%
2.70% 0.00% 0.00% 3.95% 0.45%
Healthcare 3.13% 3.35% 3.00% 0.00%
5.41% 3.77%
3.38% 0.50% 2.66% 3.49% 3.50% 3.54% 3.18% 3.02% 2.73% Seekers (Rs 200000 - 500000)
20.00% Ayurveda siddha doctor
15.57% 8.41%
18.23% 24.92% 20.29% 17.89%
17.48% 16.77% 17.94% 18.92% 18.11% 26.67% 32.43%
24.35% Education & Recreation
5.26% 29.73% 48.48%
Aspirers (Rs 90000 - 200000)
7.14%
8.13% 8.57% VHN
6.35% 11.18% 9.62% 32.73%
12.12% 15.31%
7.04% 8.24% C ommunication 15.70% Deprived - upper middle income (Rs
7.00% 8.33% 18.65%
62000 - 90000)
13.19% 6.45% 4.81%
10.63%
11.64% 36.89% 39.33% Private C linic
12.56% 1.74% 2.83% 12.09%
13.10% 14.22%
1.21% 36.67% 32.43% 27.03% 60.00% Deprived - middle income (Rs 45,000 -
2.59% 2.81%
Travel 62000)
2.98% 3.28% 13.52% 32.63%
3.26% 27.27% Private hospital
2.87% Deprived - lower middle Income ( Rs
Housing and Utilities 51.82% 22500 - 45000 )
52.56% 51.23% 48.72%
44.79% 45.70% 42.59% 41.10% 33.87% 35.77% 35.14% Government hospital Deprived - Low Income (Case Study: Building Claims Analytics solutions for Healthcare
Project Context:
Our client approached AAUM to build and implement analytical models for their Healthcare Claims Intelligence
and Insights solution that was targeted at US Healthcare providers
AAUM’s Contribution:
• A workflow and architecture for the claims engine was designed after interactions with the client
• Comprehensive collection of data from the client was undertaken to frame the input for the model
• Based on the design and data, the model was built with flexibility to be customized across client
requirements
Sample Deliverables
Workflow and architecture Claims, Provider, Billing team and The outcome of the approach was a
of the claims engine were call data were used as input model which could be customized
designed variables to build the model as per client requirements
- 53 -Case Study: Credit scoring framework for Micro Finance Institutes (MFI)
Project Context:
Our client wanted Aaum to evaluate and identify a suitable credit scoring framework for scoring MFIs
AAUM’s Contribution:
• Aaum presented a white paper evaluating various credit scoring methodologies
• Plan in progress for solution implementation.
Sample Deliverables
- 54 -Case Study: Business intelligence dashboard
Project Context:
Our client approached Aaum to build cost-effective analytical dashboards to reap the benefits of analytical
insights for their business activities.
AAUM’s Contribution:
• Customized analytical dashboards specific to client business
• Dashboard and reports developed and customized specific to roles.
• Appropriate KPIs and metrics generated specific to the client business.
• Ability to store reports in multiple formats
Sample Deliverables
Dashboard and reports were Enriched with reports, KPIs and Multiple possibilities to store
developed/deployed specific to information as per user content in the format desired
roles … privileges … by the business …
- 55 -Case Study: Design & delivery of innovative financial products for empowering SHG
Project Context:
Aaum is developing the individual credit scoring model in partnership with NABARD, National banks,
Government, NGOs to grade individuals/SHGs.
AAUM’s Contribution:
• AAUM team simplified the process with no physical requirement to travel physically to SHG/NGO locations.
• The team is devising credit score algorithms and configured in the server to perform automatic scoring.
• The entire model is getting deployed in the mobile phone to facilitate quick reach in the remote places.
Sample Deliverables
The system would captures all receipts (for the money collected from individual members)
- 56 -Case Study: Geographic Dispersion of Business Risks for a financial institute
Project Context:
Our client wanted AAUM to build geographical dispersion of business risks in India
AAUM’s Contribution:
• AAUM identified the variables needed to build risk at a location and created demographic risk framework
• AAUM used demographic risk frame work and identified risks for 60 districts in 18 states in India
Sample Deliverables
- 57 -You can also read
