Playing cupid is not an easy task but BharatMatrimony seems to have cracked the code of excelling at what they do — pairing up compatible couples for marriage. In Indian society where arranged marriages are still a way to seek for life partners, BharatMatrimony has brought quite a revolution since its inception in 1997. In an age of dating apps and social media platforms, they have been able to steal the show, thanks to data analytics.
They rely on robust analytics and advanced matchmaking algorithm to guide the members to find their life partners, enriching them through their discovery process. Leading the data science to practise at Matrimony.com is Meenakshi Variankaval who is the head of Data Science Labs. She has over two decades of experience in using data to produce actionable insights for businesses. She has worked across sectors such as life sciences, healthcare, retail and more across companies such as IBM, Cognizant, GE Capital.
Analytics India Magazine got in touch with Variankaval to understand how they use analytics and AI for the match-making process.
Analytics India Magazine: How are you using analytics and AI in the match-making process?
Meenakshi Variankaval: We use analytics to guide the users throughout their match discovery process. Matrimony’s Intelligent Matchmaking Algorithm (MIMA), which is our advanced match recommendation engine, combines machine learning techniques with mathematical rules to serve appropriate profiles to members, thereby enhancing the user experience. It combines the user’s stated preferences with their behaviour on the product to prioritise the matches for them.
Based on their viewing and communication patterns, we show sections of prospects that will lead to a higher contact initiation and mutual value creation for the users. We also keep spams in check to ensure that the user gets only relevant matches and communication, to help them with a faster turnaround in finding their desired partner.
AIM: How has analytics proved to be beneficial in the success story of BharatMatrimony? Could you highlight with few use cases about how analytics has helped?
MV: For BharatMatrimony and other matchmaking services of the group, we use data and analytics as a key enabler in all decision making processes.
- For acquisition marketing, we track the performance of marketing channels and campaigns and measure the return on investment which, in turn, is used to optimise the allocation of spends in the future
- We identify customer needs, based on their profile and activities to identify their product propensity
- We segment and profile customers based on the value delivered and identify gaps in service delivery
- We use scoring models to identify and stall suspicious user activity
AIM: When did you adopt analytics in your working? How has the revenue generation been effective since the adoption?
MV: Our analytics journey started more than a decade ago with the setting up of a comprehensive Data Warehouse. Since then we have progressed across the value chain with setting up of reporting platforms, analytics-driven campaigns, data science teams, big data platform and application of predictive and prescriptive analytics.
After the MIMA launch, we saw a significant increase in our C2C metrics and these metrics have seen a steady increase then on with continuous refinements to the algorithm. At a revenue level, the increased C2C interactions have resulted in increased customer engagement with higher user satisfaction. Over the long-run, this has been also justified by a higher percentage of payments leading to increased sales.
AIM: How are you leveraging data for the growth and benefit of the business?
MV: We use data-driven solutions to enable evidence-based decision making across areas including customer acquisition, engagement and experience; customer retention and value maximisation. This has helped us grow the business in terms of the user base, revenue and customer success stories.
AIM: How do you make use of recommendation engines? What are the various algorithms that come into play?
MV: The diversity and the numerous factors that come to play during matchmaking — ranging from personal interests, education, language, career, family, lifestyle to horoscopes is what makes it interesting for us to understand behaviour, history and more to match.
We employ data-driven techniques that offer real-time suggestions and recommend relevant profiles to our members and help them find a match based on their preferences. We have also deepened behaviour-based personalisation.
AIM: How does the analytics toolkit of BharatMatrimony look like? What are the various analytics and AI tools that you use?
MV: We use Hadoop environment for Data Lake, Python and SPSS for ML models, UNICA for campaign management and Google analytics for web analytics.
AIM: How big is your data science team? What are the various roles that you typically hire for the analytics and data science team? Also what kind of skills do you look for?
MV: The Data Sciences team has more than 15 professionals including data science professionals and Business Intelligence (BI) reporting professionals. We hire across different skills – data scientist with skills in building ML models, analytics consultants who frame and test hypothesis using large databases, and BI reporting skills to create reports to help impact the business.
AIM: What are the challenges you face while hiring analytics and AI talent?
MV: We look for people who can spot opportunities for application of analytics along with technical skills, as we run multiple parallel experiments to continuously refine the product experience for our customers. Availability of talent that possesses a good mix of technical skills (ML and AI) and product analytics experience is a bit of a challenge.
AIM: What have been the challenges of adopting analytics at Matrimony.com?
MV: Aligning our analytics roadmap to the company’s business strategy is the key to a successful implementation of the analytics program. To make this happen we strive to strike a balance between long term technical projects that are transformative in nature with quick win continuous improvements projects.
