Oleve, a startup with only four people, has achieved annual revenue of $6 million and served more than 5 million users in two years with its innovative “new lean startup” methodology. This article provides an in-depth analysis of its success secrets, revealing how small teams can create replicable success models in the AI era.
Have you ever wondered if a company with only four employees can make $6 million a year? It sounds like a fantasy, but Oleve is making this seemingly impossible story a reality. I recently delved into this startup, led by Sid Bendre, and discovered a striking fact: they not only achieved this staggering financial figure but also served over 5 million users in just two years, becoming profitable from the ninth month.
What excites me even more is that their success does not rely on traditional venture capital and large team expansion models, but has created a new “lean startup” methodology that is redefining how small teams can create great value in the AI era.
When I saw Sid Bendre’s decision to give up his position as a new graduate at Palantir, I thought at first it was a crazy move. As one of the most well-known big data companies in Silicon Valley, Palantir offers generous salaries and stable career paths. But when I learned his reasoning, I began to understand the vision behind this decision: he wanted to create a “consumer version of Palantir.” It’s not just a slogan, but a thoughtful strategic choice.
Palantir is known for its strong data processing capabilities and government collaboration, but Sid sees a huge opportunity to apply similar technical capabilities to consumer products. He and his team are not just making a product, but building a systematic machine that can continue to produce popular consumer products.
This choice made me think about a deeper question: In today’s rapid development of AI technology, is the traditional entrepreneurial model outdated? I found Oleve’s story to be more than just a success story, but a prediction of the future direction of the tech industry as a whole. They demonstrate that small teams can achieve or even surpass the output efficiency of large companies with the right methodology and AI tools. This shift not only changes the cost structure of entrepreneurship, but also redefines what it means to “scale”.
The evolution path from Quizard to the product matrix
Let me sort out Oleve’s trajectory from the beginning, because the story is full of details to learn. On January 26, 2023, they launched their first product, Quizard AI, a mobile learning app. At that time, ChatGPT had just been released for a few months, and the entire AI application market was still in the early stages of exploration. But Sid and his team showed amazing market acumen by promoting the launch of the product with a TikTok video with the concept of “what it would be like if ChatGPT and Photomath had children.” The creative video garnered 1 million views overnight and converted into 10,000 users within 30 hours.
What impressed me even more was their early technology strategy. At the time, they had almost no cost to use large language models because they found a clever way: using OpenAI’s Codex model.
To achieve these three challenges, product managers will only continue to appreciate
Good product managers are very scarce, and product managers who understand users, business, and data are still in demand when they go out of the Internet. On the contrary, if you only do simple communication, inefficient execution, and shallow thinking, I am afraid that you will not be able to go through the torrent of the next 3-5 years.
View details >
While Codex was originally designed for programming, through clever prompt engineering, they made it possible to engage in natural conversations in the open realm. They even cycle through 10 different accounts of their friends to circumvent usage restrictions.
When OpenAI finally contacted them to inform them that they were one of the highest users of the Codex model, it was both an endorsement and a sign that they needed to switch to paid GPT-3.5, which in turn pushed them to place more emphasis on product monetization.
In terms of product monetization, their strategy is also very interesting. A month later, they added the paywall feature, tried different business models, and finally opted for a subscription system. This decision may seem simple, but it reflects their deep understanding of consumer behavior: student users, despite being price-sensitive, are willing to pay for a truly valuable learning tool, especially if it can actually help them solve their learning problems.
By the fall of 2023, their street interview marketing campaign at prestigious universities was a huge success, with the first few videos being their content when searching for keywords like Harvard, New York University, Boston University, Columbia University, and more on TikTok. This marketing method not only gained a lot of exposure, but more importantly, built brand recognition.
After a year and a half of experience, they launched their second product, Unstuck AI, a learning companion tool, in August 2024. This product launch demonstrated the strength of their learning capabilities: using all the growth and product development strategies they had accumulated before, Unstuck AI reached 1 million users in nine weeks and generated over 250 million social media views in a month.
This data made me realize that they are not just making a product, but building a replicable success model. Recently, both of their products entered the top 10 in the education app rankings, with Unstuck even ranking third, behind Gauth and Duolingo.
Their latest third product is even more impressive: it is their first product outside of education, and it was developed in just three weeks, thanks to the various “blueprints” and reusable components they had previously built. Although this product is still in stealth mode, it has gained more than 1,000 users and has already achieved profitability.
This speed and efficiency of development has shown me the great value of standardizing the product development process, and when you have a mature technical architecture, growth strategy, and operational process, the launch of a new product is like adding a new model to an existing production line.
The six core principles of the new lean startup
After an in-depth analysis of how Oleve operates, I found that their success stems from six core operating principles that form the basis of what they call the “New Lean Startup” methodology.
The first principle is “either recruit the right person or not”. It sounds simple, but it requires a great deal of discipline to execute. They only hire 10x more efficient talent with multiple complementary skills. For example, their product engineers are not only full-stack developers, but also have excellent product thinking and computer networking fundamentals.
Their marketers program and designers are able to build products. The logic behind this recruitment strategy is: instead of hiring ten people who specialize in one field, it is better to hire three geniuses who can each span multiple fields. This approach reduces communication costs, improves decision-making efficiency, and allows every team member to understand the full picture of the business.
The second principle is “profit-first thinking”. In today’s environment where many startups pursue valuation and funding amount, Oleve has always prioritized profitability. They believe that profit is power and profit is focus.
This way of thinking allows them to have a clear criterion for making any decision: Can this decision increase profits? This approach may seem conservative, but in fact it greatly improves the quality of their decision-making and execution efficiency. When you know that every decision directly impacts the bottom line, you will naturally be more cautious and efficient.
The third principle is “Does this decision drive your key metrics?” “Everyone in the company is responsible for a key performance indicator (KPI), and this KPI alignment eliminates the need for micromanagement as everyone focuses on driving their own metrics on a weekly basis. This also means that all decisions must be validated against this KPI.
The ingenuity of this approach is that it breaks down company goals into personally controllable metrics, making it clear to everyone how their work directly impacts the company’s success.
The fourth principle is “continuous process improvement”. With any repetitive process, they always ask: How can we do better? What could be improved? What was wrong with the last execution? They view failures and issues within the company as systemic failures, which allows them to establish feedback loops to improve processes. This mindset ensures that they don’t repeat the same mistakes, and each iteration is better than the last.
The fifth principle is the use of “super tools”. They admit to being lazy and like to integrate as many workflows as possible into one platform. This forces them to rethink how they use existing tools.
For example, they use Launch Darkly, a feature management platform, as a manual traffic load balancer, which is placed between all large language model calls so that traffic can be rerouted to different large language model providers based on rate limiting, different strategic initiatives, or other factors. This innovative way of using tools not only saves costs but also enhances system flexibility.
The sixth principle is “don’t learn the same thing twice”. They build compound benefits by investing in technology strategies and operational blueprints. This allowed them to reuse their learnings into new products, which is why they were able to reach 1 million users on Unstuck in nine weeks—and they applied all the lessons they learned on Quizard to the new product.
This approach to knowledge accumulation and reuse allows each new product to stand on the shoulders of the previous one, achieving real exponential growth.
Organizational Structure Innovation: Reaper vs. Nurturer Model
Oleve’s organizational design reminds me of Palantir’s successful model, but they adapted it to the consumer product space, creating a unique “reaper vs. breeder” model. This organizational design not only solves the problem of resource allocation for small teams, but also ensures the balanced development of product innovation and infrastructure construction.
The “Reaper” role is similar to Palantir’s Delta engineer, but focuses on consumer products. These product engineers are the true product owners, and they are closely related to the success or failure of the product. They immerse themselves in product metrics, conduct AB testing, build features end-to-end, and work with marketing teams to effectively own the entire lifecycle of the product.
The core responsibility of a reaper is to build products that people really want and are willing to pay for. The ingenuity of this persona design is that it focuses product responsibility entirely on one person, avoiding the common problem of liability scattering in large companies.
The “Breeder” role is similar to Palantir’s Dev engineers, focusing on platform building. The main goal of these AI software engineers is to build the company’s intelligent operating system, and they advance automation in different business units such as marketing, design, product, etc., with the idea of scaling the infrastructure that affects all users and helping the company win in every market. Incubators create infrastructure that allows companies to launch and scale products faster in any market.
The innovation of this organizational structure is that it combines short-term product success with long-term technical capacity building. Reapers focus on immediate product outcomes and user value, while nurturers build the technical foundation that underpins future growth. This division of labor allows Oleve to respond quickly to market demands while continuously improving overall operational efficiency. What’s more, this model is scalable: as the company grows, they can add more harvesters to reach more markets, while breeders continue to optimize and expand the underlying system.
I find this organizational model particularly suitable for startups in the AI era. In the traditional model, product development, technical architecture, and marketing often require a large number of dedicated teams. But with Oleve’s model, AI tools and intelligent automation allow a small number of elite talents to take on tasks that would otherwise require a large team. Reapers leverage intelligent systems built by breeders to develop and optimize products with unprecedented efficiency. This model not only reduces labor costs, but also improves decision-making speed and execution quality.
Deep application enhanced by AI tools
After learning how Oleve uses AI tools, I realized that their use of AI is far beyond my imagination. They don’t just use AI as a product function, they use it as the core engine of the entire company’s operations. This all-encompassing AI integration opened my eyes to new possibilities for future company operations.
At the product level, their technical strategy reflects a deep understanding of the characteristics of AI models. While many companies are passionate about model routing, Oleve finds cue routing to be more effective. They use fine-tuned feature extractors to understand the type of user question and then route to the right prompts, tools, and examples based on the specific characteristics of the question.
For example, math problems require step-by-step answers, while history questions may require more narrative answers. The advantage of this approach is that their system automatically benefits as the underlying model improves, without the need to retrain or adjust the routing logic.
At the infrastructure level, they used some very clever technical tricks. For example, when using Azure AI Search, they encountered issues with per-storage billing. To control costs, they built a de-indexer that runs every few days to check the contents of the index and remove files that have not been used for a long time.
When the user needs these files again, the system immediately reloads. This approach allows them to pay only for content they use frequently, significantly reducing operational costs. This strategy is particularly useful for consumer products, as many users conduct “novelty tests” — trying out features they wouldn’t normally use, but which take up storage and are rarely revisited.
When it comes to marketing automation, they are building a complete intelligent system. Their marketing leaders spend a lot of time every day researching trends on TikTok to understand the types of content recommended by the algorithm. Oleve is building AI agents to automate this process, allowing AI systems to continuously monitor and analyze social media trends to identify marketing concepts that may be suitable for their products. This automation not only saves manpower but also enables the processing of data on a larger scale than humans, uncovering more potential opportunities.
In terms of product decision-making, they integrate AI into the entire decision-making process. They use intelligent systems to research new markets, identify profitable product opportunities, and even score strategic matches for potential acquisition targets. This systematic approach to market research allows them to identify which area the next product should enter more quickly and accurately.
What impressed me the most was their long-term vision for AI agents. They are building a three-phase automation system: the first phase is human-led tool enhancement, building specialized tools for team members; The second stage is workflow automation, taking over the entire process; The third phase is to integrate all workflows into a single autonomous decision-making system.
Their goal is to build a company where they hire talent for strategic insights, talent, and taste, but the entire company is run through AI agents. This vision sounds like science fiction, but based on their current progress, I believe it is entirely possible.
A scientific approach to viral growth
Oleve’s achievements in viral marketing have made me rethink what the “science of deviralization” is. Instead of relying on luck or a one-off burst of ideas, they build a system of repeatable, predictable viral growth.
Their first major breakthrough came from a deep understanding of the platform’s features. On TikTok, they found an important pattern: certain combinations of elements in videos significantly increased their viral rate. For example, when they use the post-it note concept in a video, the background is usually a MrBeast video or a Subway Surfer game footage, the angle is subtly pointed at an interesting visual element, and the handwriting must be very clear. These seemingly random details have actually been experimented and optimized a lot.
One video even performed so well that it directly pushed their app to fourth place in the education category, alongside giants like Photomath and Duolingo.
Their campus marketing strategy also demonstrates systematic thinking. In the fall of 2023, their street interview campaign at a prestigious university was a great success. When you search for Harvard, NYU, Boston University, Columbia University on TikTok, the top three to five videos are their content, and one of them gets 11.7 million views.
While this type of content doesn’t necessarily translate directly into product registrations, it builds strong brand recognition and allows users to learn about the brand through the content and then actively seek out and try their apps.
They borrowed Dupe.com’s marketing concepts when launching Unstuck AI. Dupe.com is a platform for finding similar products, particularly in the furniture sector. Oleve borrowed their post-it note concept and gained 250 million views through the concept alone in a month.
This “concept borrowing” strategy reflects their keen insight into market trends: they do not need to invent new marketing concepts, but are good at identifying and adapting proven concepts to their products.
What’s more, they systematize viral marketing. Their platform team is building tools to automate the process of creating viral content, including content monitoring, real-time feedback loops, and relationship management with creators and influencers. Their vision is to empower a strategist to command an army of dedicated AI agents rather than a team of managers. These agents perform specific tasks, and the system improves over time.
I found their understanding of consumer behavior to be particularly profound. They recognize that the consumer software market is now less like traditional tech companies and more like fast-moving consumer goods (CPG) companies, with distribution and branding at the core. Consumers now have a more refined taste in how to buy consumer software, and their experience of paying online gives them a better idea of when they want to buy or continue to subscribe to a product. This market insight allows Oleve to turn a competitive market into an advantage, where good distribution strategy and brand building are even more important.
Subversive thinking on the traditional entrepreneurial model
Oleve’s success has made me think deeply about the challenges and opportunities faced by traditional entrepreneurial models in the AI era. Their story is not only a success story, but also a prediction and validation of the future development direction of the entire entrepreneurial ecosystem.
The traditional entrepreneurial model usually follows the path of forming a large team, raising large amounts of capital, expanding rapidly, seeking more financing, and continuing to expand until it reaches some scale effect or is acquired. This model has created countless success stories over the past few decades, but it also brings problems of high cost, high risk and inefficiencies. Many startups hire in large numbers before validating product-market fit, expanding massively before they can find a sustainable business model.
Oleve’s model is completely different. They focused on profitability from the beginning, controlling team size to the minimum feasible size, and using AI tools and intelligent automation to achieve tasks that would require large teams in the traditional model. Their success proves that in the age of AI, “small and beautiful” is not only possible, but may even be a better choice. When you have the right tools, processes, and talent, teams of four can create value that used to take teams of forty or even four hundred.
The advantages of this model are obvious. Lighter cost structures, faster decision-making, more efficient communication, and more agile responses to market changes. But more importantly, this model has natural advantages in risk management. When your team is small and low-cost, you can explore and experiment for longer without having to endure a lot of pressure from investors. You can focus more on product quality and user experience than vanity metrics and short-term growth.
I think the success of the Oleve model also reflects changes in consumer behavior. Consumers are now more savvy, and they are no longer easily fooled by fancy marketing tactics or the aura of big brands. They pay more attention to the actual value, user experience and cost performance of the product. In this environment, small teams’ focus on product quality and user experience tends to be more competitive than large-scale operations of large companies.
Another important observation is that the development of AI tools is rapidly lowering the technical barrier to entry for entrepreneurship. Tasks that used to require a dedicated team can now be done by an experienced engineer with AI tools. This trend towards democratization of technology means that more small teams like Oleve will emerge in the future, challenging the market position of traditional large companies.
But I also realized that the Oleve model doesn’t work for all types of startups. Some industries that require significant capital investment, complex supply chain management, or regulatory compliance still require traditional operations at scale.
But for most consumer software, enterprise software, and digital services startups, Oleve’s model offers a thought-provoking alternative.
Ultimately, I think Oleve’s success marks a new phase in the startup ecosystem. At this stage, technical capabilities, operational efficiency, and product quality are more important than team size and financing amount.
Small teams that can master AI tools, build efficient processes, and focus on user value will gain an unprecedented advantage over the competition. This is not only a challenge to the traditional entrepreneurial model, but also a redefinition of the entire business world.
In the future, I predict that there will be more “super small teams” like Oleve who will create more value at a lower cost and with greater efficiency. And those companies that still rely on traditional models may find themselves increasingly at a competitive disadvantage if they do not adapt to this change in time. This shift will not only reshape the startup ecosystem, but also change our fundamental perception of how businesses organize, work, and create value.
