Our Pivot

After being silent for over six months, it's time to announce an important decision. On 8/17, we officially shut down chatdesigner.ai, and we are now preparing to launch a new virtual try-on product targeted at online clothing retailers, under the company name Huhu.ai. Two months ago, the team had intense discussions about the name of the new product, and we ultimately decided to go against convention by using the company’s name instead of one related to the product's functionality. This reflects our return to our core mission and our determination to push forward. Now, I'd like to share what has happened over the past six months.

Back in February, during the third month of chatdesigner’s launch, we faced low revenue and stagnant user growth, which led to low team morale. We were at a crossroads, deciding whether to pivot. After discussions, the founding team decided to stop the bleeding by cutting most payrolls and strictly controlling the burn rate. We then ran a series of funnel tests with one goal in mind: should we continue or completely pivot? As often happens, things didn’t go as we hoped. After two months of continuous testing, we found that the average LTV of chatdesigner was far below the customer acquisition cost (CPA), and only with a 60% month-over-month growth rate could we potentially break even in a year. However, maintaining that growth would require significant investment in customer acquisition, which could lead to continuous losses. Whether these losses would result in the desired growth was also uncertain. After our Y Combinator interview in March, they pointed out that our pricing was too low, which prompted us to explore a pivot towards B2B. Although we didn’t make it into S24, I’m grateful to Gustaf Alstr?mer ???? and Brad Flora . The decision to pivot was easy, but deciding which industry to pivot to and how to do it was much more difficult. After reviewing all the corporate clients who had approached us, we set our sights on the fashion industry, specifically focusing on virtual try-on as our entry point.

Virtual try-on is not a new demand; it’s been around for over 10 years. From purely 3D approaches to today’s diffusion-based methods, new players have emerged continuously, but none have become dominant. The fundamental issue remains that the technology has not reached the threshold for large-scale commercial use. When we began researching this field, we had one question in mind: has the scaling law in this area finally been established? If the answer is yes, then we are at a turning point in this technology. We just need to outperform the state-of-the-art (SOTA) slightly to gain a first-mover advantage; otherwise, we would fade into the background. Although our team is small, we didn’t dare to put all our eggs in one basket. So, we split into two small teams: one focused on training the virtual try-on model, and the other on the relatively certain ad generation business. We weren’t confident enough to go all-in at that time, as we weren’t 100% sure we could surpass SOTA and achieve commercialization. Fortunately, the results were promising. In early July, we released the v1 model internally, and user feedback indicated that it had already surpassed SOTA, but SOTA doesn’t necessarily mean it’s ready for large-scale commercial use. The v2 model, released in August, further strengthened common use cases, making it capable of solving most issues faced by online clothing sellers. This is the model on which our upcoming product is based. Thanks to the significant progress made by our research team, this product has already been well received by many local sellers, even though it wasn’t a publicly available product. The company itself is also gradually getting back on track at a foreseeable pace.

Now, after going through our darkest hours, I’d like to share a few thoughts:

First, know who you are before choosing your entrepreneurial direction. For a team led by engineers and scientists, it’s best to choose a direction where the business case is clear but the technology has yet to break through. In this scenario, there will undoubtedly be many so-called competitors in the market, but don’t panic. The key factor in success is whether you can show your customers that your solution is significantly better than others they’ve seen. If you have a business background, don’t fight a losing battle in technology; instead, find breakthroughs from the business model. Our biggest mistake early on was pursuing a growth and operations-driven direction as a tech-focused team. That’s why we decided to name our new product after the company—to constantly remind ourselves of who we are before deciding what we want to do.

Second, doing one thing better than anyone else is more valuable than being mediocre at a hundred things. If you need to multiply this by a factor, it should be a thousand times better. During the chatdesigner period, we developed over 40 features in one month, but none were impressive, and the team was overwhelmed. Due to visible competition, we also found it difficult to raise our prices, and the difference between first and second place in the market lies in that the first can price according to the problem solved, while the second and beyond can only price according to the first.

Third, this is an era where experience is becoming less valuable. The history of Stable Diffusion and LLMs is not long, especially Stable Diffusion, which was invented in August 2022. A seasoned veteran with 10 years of research experience and a fresh college graduate stand on the same starting line. In this era, what matters are intuition, curiosity, and energy. Experience itself may even become a burden when it comes to taking bold actions. In the near future, recruitment methods will likely return to basics, where companies hire people, not just impressive resumes.

After experiencing all this, I increasingly feel that we are in an era of everyone mining. People are digging away with their shovels, not knowing what lies beneath the ore. If we strike gold, it will ultimately belong to all of humanity. The reward we receive is the ticket money for the privilege of showing the treasure first, but we don’t own any treasure. The most interesting part of the whole process isn’t the reward itself but the suspense and curiosity about what treasure lies beneath the ore.

To stay updated on our new product, please visit huhu.ai.