When we first covered Maya in 2012, it was a web platform with a stated vision “to empower women through access to information and a shared community”. Over the past years, the startup has evolved a lot and turned itself into an anonymous messaging platform helping people, regardless of gender, with on-demand expert advice in areas like health, psychology, social problems, and legal aspects. You can get information and advice from Maya’s pool of on-demand experts through SMS and app, for their premium service it takes only 10 minutes to get an answer.
Maya recently raised an investment worth BDT 4.00 crore from BRAC and aims to become your digital well-being partner. However, the most interesting thing about Maya is not the investment or its pivot from an earlier women-focused content platform to an on-demand information service, rather its enduring ability to evolve and sustain. The company now deals with thousands of messages a day and growing rapidly.
Paul Graham of YC wrote, do things that don’t scale, for Maya we have to wait to see whether they get somewhere or not but for now we are super excited about their new product and its future. More so because mobile is eating the world and messaging is the new social. To know more about the evolution of Maya, BRAC investment, and the future of Maya, we recently sat down with the Founder and CEO of Maya, Ivy Huq Russell. Find a lightly edited transcript of our conversation below.
During the early days of Maya, I was the only person doing everything with help from a couple of doctors who helped out voluntarily and I am super grateful to them. It took us almost 48 hours to answer any question in an appropriate manner. Then it became 24 hours and then 12 hours and for the last couple of years, it became 3 hours and now premium users are getting it in 10 minutes. We just launched our premium service, Maya Apa Plus, in partnership with Robi, the selling point is ‘get your answer in 10 minutes’.
Future Startup
How much has Maya evolved over the past years?
Ivy H. Russell
When I started Maya, I had little idea of what we wanted to do. I was juggling a full-time job and writing articles and publishing them to my site, maya.com.bd, which was more like a blog at the time. It was very difficult to navigate through the whole thing.
The question I originally asked when I started Maya was: where do women go when they need to ask questions about the intimate parts of their body? Although that was a basic question and the starting point for us but I really started diving deep when I left my job in 2014. That’s when I started reading about and understanding major trends and ideas and changes in the tech landscape across the world. And we decided to move from a mere website and pivoted to mobile and messaging from a web based content site. And we also turned Maya into a predominantly Bangla platform. It was at this point we were fortunate to receive a grant from BRAC to support our transformation from content portal to a mobile-first, Bangla-first wellbeing service
Our rule for experts, even these days, is that you have to answer in the same language as the question is. If the question is in Bangla, you can’t answer it in English. If the question is in ‘Banglish’ you have to answer in Banglish. We have a diverse set of experts who are capable of doing this.
Over the past couple of years, we have tried many different things. We failed in some and some of those things have stuck with us and are helping us to grow.
During the early days of Maya, I was the only person doing everything with the help of a couple of doctors who helped out voluntarily and I am super grateful to them. It took us almost 48 hours to answer any question in an appropriate manner. Then it became 24 hours and then 12 hours and for the last couple of years it came down to 3 hours and now premium users are getting it in 10 minutes. We just launched our premium service, Maya Apa Plus, in partnership with Robi, the selling point is ‘get your answer in 10 minutes’.
Initially, I used to call up my friends when there was a critical question and would request to answer. That’s why it took us so long. But now things are different. Over the past years, we have answered over 150,000 queries and the more and more data we are dealing with the more smarter our system is becoming.
We are figuring out how we route each question to the right expert which would, in turn, reduce the lead time of answering a question and will also increase the accuracy of the answers.
We are working with technologies like natural language processing to understand each question better and tag them appropriately so that we can direct them to the right expert.
We have a data scientist working from San Francisco looking at each word, Bangla, English, and Banglish and trying to understand how can we automate the process. Bangla is a sophisticated language. We have a lot of variations for a single word. If we can solve this problem using natural language processing and technology, we would be the first company to do so. And it will help us to significantly reduce response time and also to match the right query with the right answer.
We have been working hard on product end. For instance, our notification and follow-ups features are unique through which an expert can send notifications to a patient or schedule follow-up reminder. We have worked really hard to come up with these features.
We put a huge importance on the quality of answers. In order to make sure that our users are happy with answers we introduced rating on answers and if you don’t rate an answer you would not be able to ask your next question.
Our average answer rating is probably at 4.3 which is very good. When we get a 2 star for an answer, we do two things: we do a physical audit and try to find out why that happened and we also reach out to the user. We personally write this answer and send to the user even these days.
Now our experts are enjoying answering question because they are answering important questions and as well as relevant questions. Our processes and algorithms are becoming smarter all the time. For instance, our system can decide to route serious questions to a senior doctor and a not so serious question to a junior doctor. This increases efficiency.
As a company, we are very data driven when it comes to making a decision and looking at things and building solutions.
The interesting part of what we are doing is that we get heaps of anonymous data and we can tell what kind of problems are prevalent in which area, what age group is more prone to mental problems and all these issues just by looking at the questions we receive.
Which means looking at Maya data, you can understand health and legal status of many localities. Questions are helping us to understand geographical barriers and other problems people face when it comes to health, psycho-social problem and legal problem.
The interesting part of what we are doing is that we get heaps of anonymous data and we can tell what kind problems are prevalent in which area, what age group is more prone to mental problems and all these issues just by looking at the questions we receive.
Future Startup
How would you define Maya now?
Ivy H. Russell
In simple, Maya is an anonymous messaging platform that connects users to experts. We are focusing on three major verticals health, psycho-social, and legal. 75% of our queries are health related, 20% are psychosocial and rest 5% are legal questions. Psychosocial questions are growing fast, just a couple of months back it was around 12% but now it is 20% and growing.
For a country like us where doctor to patient ratio is 4000:1 and psychiatrist to patient ratio is 1: 1 million, there is a huge demand for both of these services and we want to cater to that need.
We believe that we are onto something big. We are growing consistently and our engagement rate is 86% which is quite extraordinary for an app like us. We have figured out ways to bringing people back to our app and engaging them. One example is: people ask questions when they are pregnant and continue to do so till the end of their pregnancy. We stay with them throughout the time.
Now when someone asks a critical query, we schedule a follow up within a week with them whether they are doing the things that are being suggested. We have automated check-in and we have special button designed for experts to schedule a follow-up with a user. A notification goes to the users saying that have we answered your question or have you taken the medicine that we suggested. It has dramatically improved our retention rate.
Obviously, I would say we are very proud of our partnership with BRAC. We got the investment, that’s one and then we could not ask for a better investor than BRAC. BRAC has the reach and expertise in the field we are working and moreover, they are now on our Board.
For us scale is important and we think this partnership with BRAC will help us to get to more users. In a market like Bangladesh, it is important to provide high touch support for high tech solutions, and BRAC are the ideal partner to help us do this effectively and with massive scale.
Future Startup
Maya Apa Plus sounds an interesting service but what are the upsides you see in it for the serious users?
Ivy H. Russell
Accessing doctors and experts is difficult and often expensive in Bangladesh. Here we are giving you access to experts for only BDT 60 a month for app users with unlimited questions, which is the value of internet package Robi is offering now for our premium service.
For SMS-based questions, it is important to note that we only charge once a user gets a resolution to their initial query. If the cost is even BDT 10 a day and if you get a satisfactory answer from a doctor or a lawyer or a psychiatrist, I think that is really a great option. You don’t have to pay for the consultation, brave traffic or endure any other hassle for that matter.
Now that we have come together with BRAC, they have heaps of data and access to remote areas in the country. I think together with other technology and online data, we can figure out how we can help more people and what people are looking for and how quickly we can help them.
In simple, Maya is an anonymous messaging platform that connects users to experts. We are focusing on three major verticals health, psycho-social, and legal. 75% of our queries are health related, 20% are psychosocial and rest 5% are legal questions. Psychosocial questions are growing fast, just a couple of months back it was around 12% but now it is 20% and growing.
Future Startup
How big is your team now?
Ivy H. Russell
We are a team of 15 full-timers. And currently, we have 25 on-demand experts.
This model has several upsides, we can have as many experts as we want from anywhere. Many experts who are busy can work from their workplaces and reply on their own time. Moreover, this would allow us to have people from all the regions in the country. Now anyone from anywhere can work with us.
We pay the experts based on answers. We have developed a payment module and fine-tuning it as we go. We have a training module and a separate website for the experts. Before onboarding any expert, we take them through test and training. Once they pass the test and finish their training and get vetted, we then approve them and they get access to the dashboard.
Future Startup
Are these experts known to users? Can people know who is answering their queries because it is important and it needs to be transparent that laymans' are not answering difficult queries?
Ivy H. Russell
Yes. Once an expert passes certain tests, we make his/her profile public on our website. You can see profiles of our current experts on our website.
Future Startup
How do people work at Maya? Please tell us about work culture.
Ivy H. Russell
We use Asana a lot for collaboration. At Maya Apa, we have a saying, if it’s not on Asana, either you are not working on it or it is not your work. Since I maintain a long distance office, we use Asana a lot for collaboration. We use Google docs a lot to collaborate as well and give feedback. We use a lot of different tools for data collection and analysis including firebase, tableau and few others.
We are very data driven when it comes to making a decision about which feature to introduce or not. We don’t add anything on our app if it is not backed by data. If data is telling us that this is going to work, we go and do that.
We are a team of 15 full-timers. And currently, we have 25 on-demand experts. This model has several upsides, we can have as many experts as we want from anywhere. Many experts who are busy can work from their workplaces and reply on their own time.
What are the challenges for Maya now?
Ivy H. Russell
Short term challenge for us is to build the vetted expert network which is about scaling our on-demand model. Finding great talent is always a challenge. We are hiring some people on the senior roles. We are in the process of hiring our COO. Once the C-level recruitment is done, we will go for the next phase of hiring.
Our plan is to go as lean as possible. We want to optimize at the maximum level, doing most with the limited resources that we have.
Future Startup
Now that you have raised investment, the common question is: what are the plans going forward?
Ivy H. Russell
This year, our goal is to reach millions of users, obviously hit the first million. Our projection-wise, we are moving in the right direction. We plan to expand our on-demand expert base so that we can serve our growing user base and expand our service.
After our collaboration with Robi, 600 more questions got added per day. On a good day, we receive around 1000 messages and on a bad day, we receive around 500 messages. That’s just with one premium product.
We also plan to move on to other markets. We have done research in India, perhaps we will start with Calcutta and the Middle East is also our target but nothing is finalized yet. Since we have BRAC as a partner, we may launch in countries where BRAC has operations and where there is a need for our service.
At the same time, we have kicked off our monetization side with doing the deal with Robi. that’s just the beginning. We plan to get into more partnerships in the coming days. We are in an exclusive partnership with Robi for one year. We plan to package differently for other partners.
We are also in talk with a couple of private companies who want to avail our service for their employees. Our B2B service is about to start.
We plan to invest in product development and engineering and in onboarding more experts and also training our experts given the cultural setting we have.
This year, our goal is to reach millions of users, obviously hit the first million. Our projection-wise, we are moving in the right the direction. We plan to expand our on-demand expert base so that we can serve our growing user base and expand our service.
Update: This interview has been updated for clarity.
