How Breathe Happy Leveraged Computer Vision, ML To Fix Passive Virtual Yoga Experience

Online yoga classes can be exciting and tricky sometimes — whether you are fixing a dodgy hip or mastering your first handstand, unlike physical studios or centres, where you have teachers to ensure your postures are right. Online yoga classes, on the other hand, do not have a teacher to help you with alignment, besides the fact that you have to watch the instructor carefully, avoid poses that are likely to hurt your neck, or that involves twisting too much.

If you are a person who believes in the power of yoga, but restrained from practising due to lack of a trainer who gets your posture right? Breathe Happy leverages computer vision and machine learning to create real-time insights for remote yoga teachers to train learners in the right posture and more. 

Founded by Shiti Rastogi Manghani and Radhika Dimri in 2019, Breathe Happy offers personalised virtual yoga coaching with a two-way model for trainers and learners. “Safety is the core requirement of having any sort of fitness practice done online,” said Manghani, cofounder of Breathe Happy. 

Manghani, who comes with 12 years of experience practising yoga, said that even before the pandemic, there was a huge gap in terms of how users were trying to connect with right yoga teachers and instructors. “While online yoga classes have existed for quite some time now, be it YouTube videos or Zoom and Google Hangout classes, it has always been one-way communication, and there is no engagement, feedback or motivation,” claimed Manghani.

Recently, the company also raised $300K in pre-seed funding from early-stage venture capital firm Bethnal Green Ventures and Innovate UK Grant (under ‘tech for good’ accelerator programme). With this, Breathe Happy is looking to enhance its product development and expand its engineering and marketing team in the coming months.

Breathe Happy: Yoga Meets Computer Vision 

Breathe Happy’s core USP is a computer vision model and proprietary machine learning algorithm using data sets or yoga videos. This helps trainers guide learners on the right postures in real-time.  

“With the advantage of having done thousands of classes online, we have a lot of own data sets in terms of how people of all levels — from beginners to advanced —  move in the same postures,” said Dimri. 

The videos are then run through the system which essentially tracks body biomarkers such as elbows, knees and ankles. These are recognised by the startup’s algorithm and the software then guides the users on the right posture. 

Breathe Happy claims to have aggregated all of this data across different classes and groups, and has managed to create various models that very closely mimic different postures (asanas) in yoga. “Today, we are using this in yoga to compare what the teacher’s ideal movements are, and then comparing that with students. But in the coming times, we could be doing this for all kinds of postures across fitness, aerobics, Zumba and other dance forms etc,” the cofounder added.  

Breathe Happy also said that it is focussing on building an interactive tool to enhance the experience of teacher and students relationship, where they all can come together as an online community to share their achievements and progress. In addition to this, machine vision further helps teachers spot the wrong postures quickly and correct them in real-time. Currently, in a beta testing phase, Breathe Happy has close to 50 teachers on its platform. “We are very picky about the certification and quality of trainers,” Manghani told us. 

The Money Behind Virtual Yoga 

Revenue wise, the company said it follows a subscription-based model, however, in the initial days, the team ran demos and free trials to collect user feedback and data to train the models. At present, the subscription fee for customers ranges from £40 to £20 (INR 4000 to INR 2000) per month. In terms of users, it claimed to have built a community of over 2000 subscribers with a 90% user retention rate. 

For teachers, the company said that it is looking to develop a SaaS product to enable conducting online yoga classes seamlessly. Further, it said that its teachers earn somewhere around £500 (INR 50K). However, the number of classes, price points they choose for those classes, and the number of audiences among other factors, influences their earning.  

With a monthly revenue run rate (MRR) of $1000+, Manghani told Inc42 that the startup is looking to grow this by 10x over the next year. It would launch the posture estimation model and machine learning product in February next year, which the cofounder claims would help scale the business multifold. 

“In the next three year, we are looking to get to $10 Mn by doubling down on our marketing efforts to get customer acquisition going,” she added. 

According to Allied Market Research, the global yoga market is expected to touch $66.2 Bn by 2027, from $37.5 Bn in 2019, growing at a compound annual growth rate (CAGR) of 9.6%. Further, the study stated that the offline segment still dominates in terms of market share, however, in the forecasted period (2021 to 2027), the online segment is expected to contribute the highest CAGR. 

The cofounder told Inc42 that today there are about 300-500 Mn yoga practitioners, globally, and it aims to capture 2% of that market share, eyeing 10 Mn in revenue in the next three years. “We imagined the industry to have taken five years, instead, it has actually happened in five months,” avered Manghani, pointing at the rising trend for online yoga classes amid the pandemic. “The internet search result for ‘online yoga’ has surged by 1500% and followers for yoga influencers exceeds the topics like cooking and health, combined,” Manghani added. 

Besides Breathe Happy, there have been a slew of yoga and meditation startups in the country including Jennifer Lopez, Malaika Arora-backed SARVA, Mindfit, Qriyo, Namyata, Mindhouse and others. At the same time, there exist various yoga centres run by renowned yogis and babas, including The Art of Living, Isha Yoga Center, Parmarth Niketan Ashram and others, who are also looking to provide yoga classes online. 

“We are going a step higher, where we are not only replicating the offline experience online, but we are quite objective in our approach, offering real-time insights and technology tools to facilitate teachers,” said Manghani, stating the USP of Breathe Happy.