From 200 Users to $10M ARR: The Feature Nobody Asked For

Only 200 users after three months. Then, seven months later, 5 million signups and nearly $10 million ARR. What happened in between wasn't a miracle—it was a painful pivot that most founders would have avoided.

Young Zhao, CEO of OpusClip, built what he thought was the perfect AI-powered live streaming tool. Users called it "cool" but wouldn't pay a dime. Then he discovered something that changed everything: the one feature everyone actually wanted was buried as an afterthought.

In this interview, Zhao reveals why product-market fit isn't about building what users ask for—it's about finding what they're willing to pay for before you even ask, and how to turn user feedback into a competitive moat that compounds weekly.

Young Zhao: Hi, I'm Yao. I'm the co-founder and the CEO at Opus Clip. So Opus Clip turns your long form video footage into social engaging short video clips in just one click, that saves about over 99% of your time and resources and expertise of social video production. We've gathered user signups, support and trust from over 5 million creators and business worldwide.

You can probably see that people who create social apps are really introvert, they can't socialize in real life, in real person, so they have to figure out another hacky way to socialize with people, especially I want to socialize with people when I'm not sober. I co-founded a social app company that was called Sober. So the name you can tell is that after you drink, how you want to socialize with other people, right?

We raised about $1.5 million back in 2013. However, unluckily, that was just 2 months before Snapchat was launched, we shared this very similar concept that all the friendship and all the content will disappear in a few seconds or in just a day. The company definitely failed after a year. However, I really enjoyed being an entrepreneur through that experience.

Young Zhao: I think in the early stage when people become an entrepreneur is probably driven by solving their own needs and trying to expand that solution to more people, right? I think people who want to be an entrepreneur in the early days, by nature, are people who want to impact more people, impact the society and impact the world.

So back in early 2015, I went to China with a couple of other friends from Silicon Valley. At that time, the popular trend was to copy the successful business models from the US to China, and we discovered a pretty cool business model from Toast that helped restaurants speed up everything all the way from POS to Martech to HR tech. However, it didn't really turn well because of the completely different landscape and ecosystem, but throughout the journey that we realized we are actually the helper for the local restaurant, and we should not only be a software, but also services whenever needed.

Young Zhao: So I quickly formed the talent agency focused on food review, just actually 4 months after TikTok was launched. I know that short video can redefine how people market themselves and we recognize the huge opportunity out there. I spent about 4 years expanding that agency. We had about nearly 500 creatives in the team.

After running the agency for about 4 years, we realized that all the top talents left us because all of them want to be a solo entrepreneur. It is pretty easy and intuitive for a super creative person to become a solo entrepreneur. They want to create content for themselves, not for some local restaurants.

So I thought about it, instead of managing these creatives directly in my team, why don't I build some AI power product? That can enable all of the creators, there must be an AI solution that can automate as much as possible and remove the pain for creating contents because I was running the business, I know how painful it is.

Young Zhao: All the successful entrepreneurs have some sort of altruism, they want to help others. It is pretty straightforward that they can get benefit. The benefit is not commercial, is not business, but it's more spiritual benefit from helping others.

Young Zhao: So the first product idea we came about is AI powered live streaming tool, but that was during the pandemic and everyone is working from home and watching live streaming content. So we founded the company back in January 2022, we spent about half a year building the live streaming product with all sorts of cool AI features like AI generated memes, AI populated interactions, and so on. We realized that it is not solving any real pain points.

So after 3 months of the launch, we only gathered about 200 users. So these 200 users, we actually spend a lot of effort, we want them to continuously use Open Studio to give us feedback, but none of them is willing to do so. So we pay them like $20-30 dollars a session to use our product so that we can gather the feedback. But after we stopped paying for the stipend, none of them is continuously using the product.

Young Zhao: However, the last feature we built for the AI powered live streaming product was a clipping feature, because a lot of users keep asking us that after the live streaming, we want to repurpose the content so that we can reuse for distribution. So we add a feature that during your live streaming, you can hit the button and we will add a mark and after the live streaming content is ready, we can deliver the clips for you directly. And also we use some other AI to recognize the cool moments.

That was the only feature that these 200 users liked. They say all the other stuff are nonsense. It doesn't help. How can I only use this feature? So we decided to turn the feature into a standalone product. And amazingly, in just two weeks, tens of thousands of users signed up for Opus Clip, and by the end of 2023, we had nearly 10 million ARR as well as 5 million user signups, and that was only seven months after we launched the product.

Young Zhao: The real product market fit comes from people who are really willing to pay for it. When we launched Opus Studio, we asked the user during the interviews like how much you want to pay for them. And they're too shy to give us an answer. They're like, oh, I will probably think about it or something like that, right? So you can know that they don't have any dedications or commitment to pay for the product.

But for Opus Clip, we see so many users saying like, I want to pay to jump over the queue, I want to pay to guide a higher resolution, I want to pay for more downloads, I want to pay for more credits. We didn't have the payment set up on day one, but so many users just came in to knock our door and email us and call us even because we really saved tens of hours for them.

But for the live streaming tool, we didn't save any time, we didn't really help them make more money. If you can't make more money or save time for your users, they just feel like it's a cool product. It's cool from look and feel, but not from a business perspective.

Young Zhao: In the era of AI there are definitely way more teams great at technologies than teams great at building a product market fit. They believe they have a really strong technology through like a paper, through academic research, but do the users really want it? So I think definitely start with what the user wants, stay with your user or even try to become a user and then think about whether your technology is the best solution, whether AI is the best solution.

So many AI products that can deliver over 100% of quality at times, but also really underdeliver, really disappointing you. So I think from a user experience perspective, how you manage your user expectation is really tricky. Once AI delivers some results, can user touch up on that. So that's why we spend a lot of our resources building some manual features that come after the AI first approach, which means once AI delivered the result, if user think it's 80% perfect, they can still continue the workflow from what we generate for the user.

Young Zhao: This is going to really completely solve user problem, because you are building a product, you're solving a business problem that eventually has to be 100% resolved. AI, if AI can only do 80%, you have to make sure the other 20% can be well handled by human hands.

Right now at Opus Clip, we launch a new feature every single week. That is one of our culture and that is one of the reasons users love our product. They see Opus Clip, not only the current product, but how fast we out iterate the market by launching new features every single week.

Young Zhao: Making sure you experiment extensively before launching an AI product because AI is not so controllable nowadays, the result can be versatile, can go beyond even your expectation. I see most of the AI features or AI products start with like a lab, start with a beta version, and after quite a few more months, when it is really stable, when the cost really came down, then it become part of a core product.

Don't spend too much time building an AI product. Try to launch it separately. Try to launch some small experiments and get early users feedback before cutting down the cost, before integrating that into your main product.

Young Zhao: Our product is also very unique in terms of user feedback, is that the user not only directly give us feedback through using the product, but also once they post a video from our product to social medias, we can gather the number of user views, the number of comments, number of favorites, and number of reshares. So all these what we call audience feedback is actually 10 times or even 100 times more than our user direct feedback.

And we're going to use this amount of feedback to fit into our model training so that we can judge whether we are generating a good video clip even before the user publish to the social media channels.

Young Zhao: So our vision is to enable anyone to become a creator without the effort and hassle of video editing and prompting skills. We don't want AI to remove the human authenticity. We want humans to focus and spend more time on creating an idea, create your script, film the relevant contents, and then we OpusClip will take over all the boring and tedious and time consuming work.

So there are in general for production perspective, there are 3 major stages. The first one is pre-production, where you edit a topic, you write the scripts, and you prepare the storyboard, and then it comes to the production where you film the contents, you record the screen, or you prepare the media assets. The last one is the post-production, right, where you edit the video, you assemble, you cut the scenes, and you sequence them in orders.

Young Zhao: Previously, most of the AI or technology is only focusing on a specific scene or a specific frame, but nowadays, the overall story, the overall narration and the curation can be well enhanced and automated by AI. The second thing is definitely the generative AI technologies. We can see that most of the time it's really hard to film the content, but nowadays with Sora, with Pika, with other AI models, you can generate like a few seconds long footage or B-roll from just a few words.

So I think one is the understanding, the other is the generation. These two huge advancement of AI is actually automating the whole process and also generating new footage. I think with the advancement in these AI technologies, creators are actually shifting from creating content to being creative. They don't have to spend so much time to specifically create each piece of content. AI can do this job for them, but they should spend probably 10 times more time to think about how to be creative, because everyone is being more and more creative. Everyone has more time to be more creative. Everyone has more resources to be more creative.

Young Zhao: I think the paradigm shift is, how can you be 10 times, 10-fold more creative than before? I think in the AI era, being authentic and being genuine is probably going to be more of a rare find, because it's easy, it's super easy to leverage AI to generate a lot of fabricated content.

So my advice would be, just be yourself, be human, be real, and maybe nowadays it's easy to be like that, but think about 5 or 10 years later, when AI is everywhere, when AI generated content is everywhere, like being super real, being super genuine and authentic is probably going to help you be the most creative.