Many accelerators can negatively impact startups. What distinguishes the effective ones?

      Every accelerator provides a similar offering: capital, mentorship, a network, three months of support, and significantly improved survival odds. However, evidence indicates that much of this may not be effective.

      In April, Youn Baek and Deepak Hegde from NYU Stern published a working paper through the National Bureau of Economic Research, analyzing nearly 750,000 American startups across 329 programs. They discovered that between 60 and 80 percent of accelerators leave the companies that join worse off than if they had not applied. Conversely, a smaller group significantly increases funding, growth, and exit rates, including Y Combinator, Techstars, and Endless Frontier Labs.

      While the study identifies which programs are effective, it does not clarify why. To gain insights into this, we consulted Yann Goarin, a founder and expert in product-market fit.

      Goarin spent ten years at Google and YouTube, where he launched over twenty products across Europe and the U.S., and has since led product and marketing initiatives for several venture-backed startups. In 2023, he founded Zag Labs, an advisory firm that has assisted over a hundred early-stage companies in reaching the market and achieving product-market fit. He created the “PMF System,” which views product-market fit as a problem-solving process rather than a standalone event. Currently, he serves as Founder in Residence at AAXIS, guiding the enterprise technology firm’s venture-building efforts, while also mentoring at five accelerator programs across the U.S. (Techstars, gener8tor, FoundersBoost, Expert Dojo, and USC’s Iovine and Young Academy), providing him with a unique perspective on how different programs support their founders.

      Most accelerators take equity in exchange for funding and support, with their returns dependent on a few companies in each cohort achieving significant success. They offer founders leverage in multiple forms: capital, introductions to investors and customers, brand recognition, and expertise.

      Similar to top universities, the best accelerators attract and select outstanding founders. However, the chances of success remain very low. Creating a defining, venture-backed company is exceptionally challenging, with luck and timing playing significant roles. Nonetheless, there is a systematic approach involved in this process, which Goarin argues is either poorly taught or not presented comprehensively.

      Research indicates that knowledge is the most crucial form of leverage. Susan Cohen, Benjamin Hallen, and Christopher Bingham, who studied the original American accelerator programs, found that when accelerators do benefit their companies, the main factor is the knowledge those companies gain within the programs. However, this knowledge is difficult to scale.

      Goarin recalls one engagement with a seed-stage AI startup that had developed a video production platform. The founders came from one of the world’s most selective accelerators, secured $4 million in funding, and in a year reached $1.2 million in annual recurring revenue. However, they faced a churn rate exceeding 30 percent. Their approach involved aggressive selling and rapid feature development based on customer requests, supported by investors who prioritized revenue numbers.

      What the founders overlooked was that their target segments (small marketing agencies, independent video creators, and boutique production companies) did not form a cohesive market. Though they appeared to want faster and cheaper video production, their needs varied significantly in terms of production volume, quality standards, integration into workflows, and distribution methods. The product attempted to cater to all three segments but ultimately failed to meet any of their needs effectively. Subsequently, customers left faster than they could be replaced. After downsizing and pivoting, they were unable to secure further funding and ran out of operational time.

      Goarin joined the situation too late to make a change. “I expected founders emerging from a prestigious program to be better at testing assumptions and diagnosing problems. I was mistaken. They were just as lost as many of the others I advise.”

      During that period, he began mentoring at Techstars, identifying a chance to tackle these issues at scale. From within the program, it became evident how knowledge effectively reaches founders and what often falls short.

      Programs that do provide teaching tend to offer fragmented knowledge: one expert covers product, another addresses sales, and a third aids with fundraising. Founders are then left to piece it all together into a functioning company, and most fail. This approach seems peculiar from an external viewpoint. Accelerators and venture funds invest considerable resources in selecting applicants, screening thousands to identify the few worth supporting, and then simply hope they can figure it out.

      What remains unaddressed is the concept of product-market fit itself—the proper integration of these fragmented components that eventually leads to substantial demand for necessary products delivered profitably. There are two reasons it is not integrated into curricula. Product-market fit is not recognized as a distinct discipline, meaning there is no established body of practice for teaching it. Additionally, the mentorship model typically recruits subject matter experts by their specific function, leaving product-market fit, which overlaps several functions, without a designated owner—until now