Unfair advantage

How Swupnil’s unique background helped him create a product that no one else could

After receiving $50k from tennis legend Andy Roddick, Swupnil Sahai tried to pitch to VCs. “That was really brutal,” he recalls. “It seems very unlikely that you will be able to solve this problem,” they told him. “Also, tennis feels like a small niche sport. Only rich country club guys play tennis.”

Little did they know that Swupnil eventually solved that problem. And years later, VCs ended up investing in a Series A at a much higher valuation. So how did he go from working on SwingVision as an unpaid side project to becoming a viral tennis sensation?

A set of fortuitous coincidences

What’s interesting about Swupnil’s story is that he never planned to be an entrepreneur. “I’m pretty introverted and risk-averse,” he notes. “Starting a company has never been part of the plan.” Still, a series of fortuitous coincidences led him on this path, almost like life already had a plan for him.

Everything started while watching professional tennis. “I really love seeing all the stats that they’d show,” he says. Swupnil started thinking that analyzing his own stats would be a massive advantage in improving his game. However, no one had built something like SwingVision yet.

But when the Apple Watch came out, he realized that he could play tennis while wearing "a computer on his wrist." He began to imagine manually keeping score and tracking stats using the Apple Watch. With that, the first iteration of SwingVision for the Apple Watch was born.

I personally use the Apple Watch app myself, and I seriously can’t imagine playing tennis without it ever again. Still, Swupnil knew that he couldn’t build a long-term, defensible brand out of it. Most engineers would be able to spin up an app to manually track tennis stats. For this reason, Swupnil decided that SwingVision was going to remain just that — a funny, unpaid, side project.

💡 What’s your moat? While discussing this, Swupnil and I started to chat about the importance of building a strong moat around our startups. “I’m seeing a lot of ‘AI companies’ that are basically just building on top of OpenAI,” he notes. “They are essentially just calling an API. And if you’re calling an API, I can just call an API too. If you’re trying to build an enduring company, then you should focus on building your own models and collecting your own proprietary data. You should aim to be OpenAI rather than use it.”

However, after releasing the app, more and more users started giving feedback. "It would be great if it could automatically keep the score for me," they said. Swupnil didn’t believe that was even possible. “I would love to have that too,” he thought. “But it sounds ridiculous. How would you even build that?”

But after finishing his PhD in statistics, he started working at Tesla in their autopilot team. “I was really fascinated with what they were doing,” he recalls. “I learned about computer vision and I was like….Shit!!! I could use the camera to automate the scoring! This problem can actually be solved!”

Early fundraising challenges

His work with computer vision at Tesla, his expertise in statistics as a professor at UC Berkeley, and his love for tennis gave Swupnil the ultimate unfair advantage to create this exceptionally complex product. But although Swupnil had the perfect background to meet this technical challenge, the complexity of the project made seed funding difficult.

Swupnil received $250k from former no. 1 in the world Andy Roddick and a few other angel investors. However, pitching to VCs did not go as well. They just didn’t believe Swupnil was able to build the product. In hindsight, that was a blessing in disguise. First of all, Swupnil didn’t give up as much equity as he would have if he had raised funds earlier. Instead, VCs ended up investing years later (in 2023) at a much higher valuation.

Second, being undercapitalized and working without a salary taught the team the importance of unit economics. They learned to be profitable well before the recent funding crunch. Swupnil explains, “It was good because we've always been really disciplined with our burn, our hiring, and our expenses in general. I think we've come at this from a very different mindset because we were always undercapitalized. I think that's made us a healthier business.”

💡 Did you know? Andy knew how important and game-changer SwingVision was going to be for tennis players around the world. And he was right. Institutional investors might not always get your vision and your product immediately. But if you have users who love your product, you can let them invest in your startup via a community round. On Wefunder, you can raise up to $5M from your customers and fans, on your own terms, and from people who deeply care about what you’re building.

The perfect viral loop

Given the social nature of the game of tennis, SwingVision has a head start in terms of its product-led growth. Anyone using the app to record a match on the tennis court automatically becomes an advertiser, generating a powerful viral loop. The key to SwingVision's growth has been to use social media marketing and product features to make this viral loop even more effective.

First of all, they created product features that make users genuinely excited to share their recordings:

  • The app automatically removes the match’s “dead time” — when you export a video, you only get the shots where you are playing a point.

  • You can export AI-generated highlights. They show the best points, longest rallies, and service winners as well as many other pre-built videos.

  • The video includes a heatmap of the shots played and stats of your shots’ type and speed.

  • They built a feature that turns landscape videos into vertical ones. It does this by automatically "following" the player as he moves on the court. No editing required. You can export reel-ready videos right from the app.

Trust me, when you see yourself in these types of videos for the first time, it’s a truly unique feeling. Every time I shared the highlights of my matches with my friends, they were immediately hooked and downloaded the app.

The other component behind SwingVision’s growth is social media. If you play tennis, it’s almost impossible that the algorithm has never shown you a video tracked with SwingVision. I keep seeing them all the time. SwingVision users organically post them online and the company reposts them from their accounts. Plus, they also take videos from famous tennis players, apply SwingVision stats on top of them, and re-upload them from their page. It’s a truly unique, evergreen strategy for organic content marketing.

Finding the right model for product-led growth

One of the most important aspects of product-led growth is picking the right model. First of all, you will have to decide whether to go the freemium way or the free trial one. Then, you have to experiment to find the right setup in each model. We have seen this when we interviewed Luca Mastella. Adopting a freemium model without needing a credit card has been a game-changer for Learnn.

SwingVision has a combination of both a freemium version and a free trial. At sign-up, they push the user to go for a free trial (which most users do). But if they skip it, they can still enjoy their freemium version and experience the quality of the product. A game-changing experiment for them has been to test how many days of a free trial would lead to higher conversions.

Compared to most apps, where you can just download the app and play with it right away, SwingVision requires some additional time. That’s because after you download the app, you actually need to wait to get on the court to play tennis. That’s why they tested changing the free trial from 14 to 30 days. Swupnil explains, “We wanted to include that in the time of the trial. So that's why we have a longer trial than most apps, and we did see a big bump in conversions when we switched from 14 to 30 days.”

Another important A/B test has to do with the subscription plan offered. They tested offering just the yearly subscription at sign-up versus offering both a monthly and a yearly one. Offering only the yearly subscription led to a lower conversion rate. But, it also led to a higher renewal rate and a higher lifetime value. This is probably because users have more time to experience the true value of the product.


Because of Swupnil’s unfair advantage and his commitment to developing this technology, no one has been able to replicate what he and his team have accomplished. Their unique expertise and proprietary data give SwingVision a moat. Swupnil encourages entrepreneurs to rigorously build a company that doesn't rely only on OpenAI. “If you’re trying to build an enduring company, then you should focus on building your own models and collecting your own proprietary data. You should aim to be OpenAI rather than use OpenAI,” he says. “Pick a niche that you’re trying to work in and be the best person for that and have the best data for that because OpenAI won’t be able to compete with that.” If you want to follow along, you can connect with Swupnil on LinkedIn.