Many accelerators tend to hinder the progress of startups. What distinguishes the effective ones?

      All accelerators present a similar offer: funding, mentorship, a network, three months of support, and significantly improved chances of success. However, evidence indicates that much of this may not be effective.

      In April, Youn Baek and Deepak Hegde from NYU Stern released a working paper via the National Bureau of Economic Research, analyzing nearly 750,000 U.S. startups across 329 programs. They discovered that between 60 and 80 percent of accelerators leave participating companies worse off than if they had never applied. Conversely, a smaller group, including Y Combinator, Techstars, and Endless Frontier Labs, significantly improves funding, growth, and exit rates.

      The study identifies which programs are effective but doesn’t clarify why. To explore this, we consulted Yann Goarin, an expert in product-market fit and founder of Zag Labs. With a decade of experience at Google and YouTube, where he launched over twenty products in Europe and the U.S., he has since led product and marketing in various venture-backed startups. In 2023, he founded Zag Labs, advising over a hundred early-stage companies on market entry and accelerating their paths to product-market fit. His developed “PMF System” approaches product-market fit as a problem-solving process rather than a mere event or feeling. Currently, he serves as Founder in Residence at AAXIS, overseeing their venture-building initiatives and mentoring across five U.S. accelerator programs (Techstars, gener8tor, FoundersBoost, Expert Dojo, and USC’s Iovine and Young Academy), providing him with a distinct perspective on how various programs assist their founders.

      Most accelerators take equity in return for funding and three months of support, relying on a few companies in each cohort to secure significant funding or exits for their returns. They offer founders several forms of leverage: capital, connections to investors and customers, brand recognition, and expertise.

      Similar to elite universities, top accelerators attract and select the best founders, yet the success rates remain quite low. Creating a venture-backed company that defines a category is exceptionally challenging, heavily influenced by chance and timing. However, it is not purely a matter of luck; Goarin asserts there is a method to this chaos, which is either poorly understood or inadequately taught.

      Research indicates that knowledge is the critical form of leverage. Susan Cohen, Benjamin Hallen, and Christopher Bingham, who investigated the original American accelerator programs, found that where accelerators do enhance their companies, the key factor is what those companies learn during their time in the program. Unfortunately, this aspect is also the most challenging to scale.

      Goarin recalls one instance involving a seed-stage AI startup with a video production platform, whose founders came from a highly selective accelerator. Although the startup raised $4 million and generated $1.2 million in annual recurring revenue within a year, it faced a churn rate exceeding 30 percent. The response was to aggressively sell and rapidly develop, adding features based on customer requests, supported by investors’ belief that revenue was the primary focus.

      What the founders failed to recognize was that the three target segments (small marketing agencies, independent video creators, and boutique production companies) did not form a cohesive market. Though they seemed to demand quicker and cheaper video production, their varying production volumes, required quality, workflow integration, and distribution methods led to a product that inadequately served all three. As a result, customers left faster than the startup could replace them. After cutting half their team and pivoting, they could not secure additional funding and ran out of resources.

      Goarin arrived too late to influence the outcome. “I assumed founders emerging from such a program would have better testing abilities and problem diagnosis. I was mistaken; they were as uninformed as many of the others I advise.”

      Around that time, he began mentoring at Techstars, where he saw a chance to tackle the issue on a larger scale. Within a program, it becomes evident how knowledge effectively reaches founders and what often goes unaddressed.

      Programs that do teach often do so in fragmented ways: different experts cover specific domains, leaving founders to piece together a functional company. Most fail in this endeavor. Observing this from the outside is perplexing; accelerators and venture funds exert tremendous effort in selecting candidates from thousands of applications only to rely on hope that these founders will figure things out.

      Notably absent from the curriculum is product-market fit itself—effectively integrating these fragmented elements to create a product that meets a genuine need profitably. There are two reasons this is often overlooked. First, product-market fit is not recognized as a standalone discipline, leading to no established body of practices for teaching. Second, the mentorship model typically involves subject matter experts focused on specific functions, leaving product-market fit—an interdisciplinary concept—without an assigned owner, until now.

      Goarin's teaching connects all aspects of the process with a clear goal: to avoid creating something unwanted. "Accelerators provide access