Insights around building Bharat Matrimony and Caratlane.
Conversation with Sundar Ganesan, Director of Product @Caratlane, Ex Matrimony.com
Hey there 👋
Dharmesh Ba here. If you enjoy reading my newsletter and are interested in collaborating with me, check out 1990 Research Labs. We are a boutique user research consultancy located in Bangalore. We have collaborated with several top startups, including Duolingo, CRED, Plum Insurance, Udaan, Allen Digital, and Jar.
I discovered Sundar on Twitter and met him IRL two years ago. Our lengthy conversation was filled with a treasure trove of insights. Unfortunately, not all of it could be shared, but I convinced Sundar to chat with me for TIN.
Sundar started as a Product Manager in 2009, even before the role of PM became popular. He began his career with MingleBox, an Edtech platform that helped students prepare for competitive exams like the CAT. At its peak, MingleBox was a competitor to Pagalguy. He later became the first product manager at Bharat Matrimony, where he built the product team over five years. Currently, he is the Director of Product at Caratlane. Throughout his journey, Sundar has worked on products associated with social status, such as education, weddings, and diamond jewellery, which is what makes this conversation exciting.
Building Mingle Box
You started your journey with Minglebox as a Product Associate in 2009. Can you tell us what Minglebox did? It looks like the site doesn't exist anymore.
Minglebox underwent several phases since its inception in 2006. Initially, it aimed to compete with Orkut as a social networking platform. The team noticed a surge of activity within college campuses, similar to Facebook's strategy of campus networking. Consequently, Minglebox shifted its focus to providing intranet services for campuses.
However, realizing that this approach was not yielding the desired results, Minglebox pivoted to become an EdTech company around 2008-2009, which is when I joined the team. Although I did not witness the first two iterations of Minglebox, I was aware of their robust features, thanks to the platform's social and campus networking roots. With the transition to an educational platform, our task was to digitize physical examination papers, coinciding with the CAT exam's move to an online format in 2009. This shift gave us a competitive edge as we began converting numerous physical test papers into digital exams.
Our success grew as we expanded our offerings to include other MBA entrance exams, IIT JEE, GRE, GMAT, and more. Minglebox was at the forefront of technology, launching one of the first Android apps in the education sector around 2010-2011. Later, the company was acquired by a Chennai-based firm, Quscient technologies. This company focused on assisting students with studying abroad, primarily through lead generation services.
Minglebox continued to operate until around 2017-2018. Since then, I have not kept up with its developments.
Did the term product management exist back then in 2009? What was your understanding?
When I joined the company, my first position was as a Product Associate. Within a year, I was promoted to Senior Product Associate. At the time, I didn't understand the difference between the two roles. Another year passed, and I became a Product Manager, taking on multiple responsibilities. Since the core product we were working on was CAT preparation, and I was also preparing for the CAT exam, I began to approach the product from a user's perspective. I would discuss it with friends and even took the CAT exam sponsored by Minglebox to gain firsthand experience.
Back then, the CAT exam spanned over 30 days, not just one. Candidates were eager to discuss each day's paper, the difficult and easy questions, and the types of questions asked. In response, we aimed to build a community around this interest. There was an existing community called Pagal Guy that was very popular at the time. We wanted to create something similar to engage our users daily and increase our daily and monthly active users. Although I wasn't familiar with these terms then, I knew we needed to keep users engaged throughout the 30-day period of the CAT exam.
To achieve this, we developed forums and a question-and-answer product. Users could post questions and receive answers from experts. This is how we built our product, working with a very small team over the next six months to a year.
Building an Android app in 2011
What were some of your interesting observations while building Mingle Box back then?
A lot of interesting insights emerged from our experience. The first insight regarding why we needed to build an Android app was influenced by the fact that in Bangalore—much like in Chennai—people spend 2 to 2.5 hours commuting to colleges on the college bus. This was around the time when Android phones were becoming popular. People began playing games like Talking Tom on these devices, and Android phones started to gain fame among college students. This trend picked up around 2011-2012.
At the time, we conducted what I now recognize as focus group studies, although I didn't know to call them that back then. I was simply visiting colleges to talk to potential students and understand their habits. They were aware of our website and the digital papers we offered. They told me that they would find it useful to have quiz questions to prepare for the CAT examination during their commute, the response was positive. This feedback sparked the idea of converting our question papers into a set of 10 daily questions.
This concept led to the creation of the first version of our Android CAT prep app. We transformed our content into a daily set of 10 questions. The most memorable part of this user experience journey was witnessing the app's growth. Initially, we celebrated milestones like 5,000 and then 10,000 downloads. While these numbers seem modest today, they were significant achievements at the time, considering Android's user base was only around 1 million to 10 million customers.
What was the business model for Mingle Box?
We offered free question and if users desired a proper mock test, they would have to pay. In another version, we proposed that the learning material would be free, but to access the questions, payment was required. This was the essence of our freemium model. Anyone could use the app for free, but taking the tests necessitated a payment.
Payments through post office money order
Did people have credit cards back then? How did they make the payments?
One day, a person from Arunachal Pradesh called up, expressing a desire to make a payment. They were intrigued by our app, and upon inquiry, they revealed how they discovered it: their father had gifted them an Android phone, and while exploring it, they stumbled upon our app. The father wanted them to prepare for the CAT exam, and since there were no CAT coaching institutes in their area—somewhere near Itanagar—they found our app particularly interesting.
You wouldn't believe how we processed the payment—it was through a money order. We instructed the customer to go to the post office, send us a money order, and I vividly recall receiving it from the postman at our office. We celebrated that first payment of 1,200 Rs, a milestone for our product. Once we were notified of the money order, we provided the customer with a coupon code, which they used to access and start using the product.
One aspect of my job that I thoroughly enjoy is connecting with customers. I often engage in random conversations to understand their thoughts on our product and their general habits. I look for inspiration in the products they use and consider how I might incorporate those insights into our offerings.
Building Matrimony.com
How did your Mingle Box experience help in setting up the product team at Bharat Matrimony?
When the CEO and the CPO of Bharat Matrimony traveled to Delhi to meet the MakeMyTrip team, they learned about the existence of a product team there. Curious, they inquired about the team's role and were told that the team was responsible for gathering requirements, among other tasks. This revelation led them to the decision that Bharat Matrimony also needed to hire a product team. That is how the product team at Matrimony came into being, and I became the first Product Manager.
What was the first thing you did at Bharat Matrimony?
There were a lot of metrics in place. We knew what we were doing, etc., but we were not able to link much of that to the product. One of my first projects was to unify all of these disparate elements. On the Bharat Matrimony website back then, users could send messages, chat, and send emails, among other activities. I aimed to integrate all of these functions, which was a significant challenge because it required changing the entire architecture. However, one positive outcome was that all metrics were now centralized. You could simply look at one dashboard to understand the health of the website and assess how the metrics were performing.
A/B testing ‘Express Interest’ button
Can you give me an example of how the metrics improved?
The CTA to connect with a person on the platform was titled "Express Interest." We started questioning what "Express Interest" actually means. You express interest to someone, but you generally don't click on "Express Interest." So, should we change it to "Send Interest"? "Send Interest" seemed like a better CTA than expressing interest, even though both performed the same action. So, we changed it to "Send Interest." We began to see some improvement in numbers. We didn't have fancy A/B testing tools back then, so we used our multiple websites to experiment. With about 300 websites at our disposal, we decided to test the new CTA on one website for a week to see if there was an increase or not. Then, we would roll it out to a few other websites.
We continued to refine our approach. Why just "Send Interest"? Could we simplify it to "Interested?" with a "Yes" or "No" response? This provided a clear action: "Yes" to send interest, and "No" to indicate disinterest in the profile. Then we questioned the need for "Yes" or "No" at all. Why not use a tick and cross, similar to the system Tinder was using?
This small CTA change led to a significant impact. And it all happened over the course of two to three years.
You said there are 300 websites. Can you throw some light on it?
Sure. The parent company is Matrimony.com, which operates two main portals: Bharat Matrimony and Community Matrimony. Bharat Matrimony caters to region-specific or language-specific demographics, offering services such as Tamil Matrimony, Telugu Matrimony, Kerala Matrimony, Hindi Matrimony, and others. Community Matrimony, on the other hand, focuses on community-specific websites, featuring options like Agarwal Matrimony, Chettiar Matrimony, Vaishnav Matrimony, among others.
In addition to these, we offered other services: Assisted Matrimony and Elite Matrimony. Assisted Matrimony provides a relationship manager who helps clients find a match if they, or their support system, are unable to do so on their own. Clients provide their requirements to the relationship manager, who then shortlists potential matches and maintains regular contact throughout the process.
Elite Matrimony operates similarly to Assisted Matrimony but is tailored for a specific clientele. It caters to individuals who prefer not to publicize their profiles on websites and seek matches among the elite, often including some of the wealthiest people. This service offers a discreet and exclusive matchmaking experience.
Men vs women in Matrimony websites
Can we talk about how the behaviour of men versus women was different on this platform?
Due to the ratio of people on this platform, metrics tend to be very different. It's not a one-to-one ratio website; there are significantly fewer women's profiles than men's profiles. From my perspective, a man is likely to send interest to all possible women on his list to increase his chances of acceptance. Instead of selecting a limited number of people, he maximizes his opportunities by reaching out to more individuals.
On the other hand, women experience the opposite situation. With their numbers being fewer, they are inundated with requests. Upon registering, a woman might receive 250 to 300 responses or an even higher number, which can be overwhelming. In contrast, a man might not receive any response at all, even after a week or ten days, failing to secure even one acceptance from a female counterpart.
In response to these dynamics, we conducted extensive research on what constitutes an ideal men's profile and an ideal women's profile. We then developed an optimal journey for each group. For instance, we found that annual income is crucial for men, so we made it mandatory for men to disclose their annual income because our data showed that women only expressed interest in men who had listed their income. It's important to note that a woman initiating interest carries far more weight than her merely accepting a man's interest. It's comparable to the difference between a woman proposing to you versus accepting your proposal; the former is considered a more significant gesture.
Pre-Jio vs post-Jio
Did you see a difference in behaviour between the pre-Jio and post-Jio launch?
We saw a significant change in how people behaved before and after the advent of JIO in 2016-17. Prior to 2016-17, people would gather together, especially on Sundays, which were our peak metrics days. Families would sit in front of their large desktop computers and collectively browse profiles, selecting interests with a yes or no. Sons, fathers, mothers, daughters, and other family members would all participate in this activity.
However, after JIO's introduction and as people began downloading our app, we noticed a substantial increase in app usage. Nowadays, I believe about 95% of our traffic comes from the app, with desktop traffic dwindling significantly. One advantage of reduced desktop traffic is the seriousness of its use. On a desktop, actions are more deliberate compared to the casual browsing on mobile. On mobile, users multitask and may switch between apps, such as responding to a WhatsApp message. In contrast, desktop users of the past dedicated specific times for their searches, treating it almost like a religious activity. They would even schedule it, for instance, from 11 to 12 o'clock, with the goal of finding a match or sending an email.
Another interesting point is the payment process on our website. We've always prioritized assisted selling, to help with payment failures. At Bharat Matrimony, a payment failure was considered an inauspicious event, given the importance of the transaction. To combat this, we employed multiple payment gateways and rerouted transactions to ensure success. We even sent notifications on auspicious days like Amavasya, encouraging users to make their payments. We knew the optimal times to send these reminders and the right people to target.
Do you have some insights about how things vary between different geographies within India?
Bharat Matrimony had two competitors: Shaadi.com and Jeevansathi from the Info Edge Group. Jeevansathi performs very well in North India, while Shaadi excels in Gujarat, Maharashtra, and among NRIs. We have historically been strong, and I believe we continue to be, in the southern and eastern regions. This success is largely due to our understanding of the nuances within these communities.
How did you solve expectations in partner search during marriage?
Many people don't know what they want on registration day, and they fail to update their preferences later, leading to seeing same profiles. We try to prompt them to consider if their preferences have changed. This is where the intelligence comes into play. When we notice that someone has expressed interest in many people who don't match their stated preferences, we ask if they'd like to adjust their criteria. This allows the machine to learn and improve its recommendations.
As people age, they often relax their requirements. For instance, we might suggest to someone over 30 to consider a wider range of heights. A person who is 6 feet tall might have initially sought matches between 5’9” and 6’, but we might ask if they're willing to include shorter individuals. This could reveal 500 more potential matches they haven't seen before.
While users maintain control over their preferences, many don't bother to change them, primarily because they're unsure of what they want at the start
Have you seen any interesting observations when it came to sending interests?
Yes, so there's a humorous story from around the time Trump became president. That's when he announced his intention to ban people from traveling into the United States, among other things. Consequently, Indians were quite concerned at that time. For a brief period, we observed a significant decline in the level of interest Indians showed towards NRI profiles on the platform.
Building Caratlane
I want to switch to your Caratlane journey. How did that transition happen?
I was at Bharat Matrimony for a little more than five years. I felt I had accomplished a decent amount of what I set out to do and was in search of something interesting. That's when Caratlane came along. Initially, I worked on the Titan Eye+ product, managing both the product and its business. When Titan Eye+ was reintegrated with the Titan team, I took on the role of managing the omnichannel product for Caratlane, along with all the new initiatives.
Solving for trust
How did you solve the issue of trust in buying jewellery online?
Our jewelry buying experience is quite unique. Typically, people visit a jewelry shop to decide what they want to purchase or examine. However, our approach is different. Customers have likely browsed the product on our website. Nowadays, they can even book an appointment at the store and request to view a specific item, say, in HSR Layout. We transfer the product within a couple of days and notify the customer that the item has arrived, inviting them to come and try it. This exemplifies how we've leveraged technology to enhance the omnichannel experience.
The second aspect is trust, which is where Tanishq comes into play. We refer to our collaboration as a Tanishq partnership, promoting it as the "Spirit of CaratLane and the Trust of Tanishq." Trust is crucial, particularly in the jewelry business, whether it involves gold or other materials. Gaining trust is challenging, but once established, it's significant. Tanishq has built confidence through innovations like the carat meter. Customers recognize that we are a 100% Titan company, and even when we were a subsidiary, the association with Tanishq and the Tata Group was clear. I often mention that the Tatas have a certain reputation—consider the widespread use of Tata salt, for example.
As a final question, what are some of the interesting observations from Caratlane?
I used to jokingly tell people that I first encouraged them to do an MBA, then to get married, and now I'm helping them choose jewelry for their wedding. My experience working with the omnichannel product was beneficial. I visited many stores—about 40 to 50 across 12 or 13 different cities in one year—spending time with our jewelry consultants and customers. The experiences were incredibly varied.
For instance, I recall a fascinating incident from last Dhanteras. A customer from the US wanted to buy something for his wife. He had selected two products and was in the process of purchasing them. His father and mother were present, and during billing, his father asked if he had bought something for his mother as well. Caught off guard, he decided to buy a gold coin, considering it auspicious to do so on Dhanteras. The emotions involved in these transactions are profound.
Many people who enter our store are celebrating significant occasions. Whether it's for a birthday, an anniversary, or another special day, they don't mind spending two or three hours in the store. The entire experience is crucial, from the moment they park their car to when they leave. Our Google Maps page features reviews praising the parking facilities. While it may not seem directly related to jewelry, the entire customer experience matters.
I fondly remember my early childhood visits to GRT, where they would offer me juice. That was part of the experience, not just about the jewelry. There's a lot of emotion involved in jewelry, but it's also about the overall experience. We're focusing on enhancing the emotional aspect of jewelry shopping. We've taken some initial steps, but there's no denying that emotion plays a significant role in these experiences.
(Fin)
What are some of the exciting insights that you have about building status-driven products?
Great insights!