Overview
Torch began in 2017 as a service that matched executives with coaches, then wrapped the relationship in 360-degree feedback, goals, and reporting software. It expanded from senior executives into a broader enterprise learning platform, merged with Everwise, and raised at least $75 million across its 2019 Series A, 2021 Series B, and 2022 Series C.[1][2][3]
The record does not support a shutdown story. YC labels Torch inactive, yet Torch's site still sells coaching, its board installed a new CEO in 2024, and current product material remains online.[4][5][6] The useful retrospective therefore examines strategic pressure. Torch proved that human coaching could become enterprise software, then entered a market where well-funded rivals began pairing human coaches with cheaper, always-on AI.
Founding Story
Cameron Yarbrough and Keegan Walden met while studying counseling psychology at the California Institute of Integral Studies. Walden had spent about six years at a large financial institution before starting the master's program. He remembered meeting Yarbrough in a seminar on gestalt therapy, training beside him, and completing the clinical internship year together. Their paths then split: Yarbrough practiced couples therapy and built the Well Clinic, while Walden pursued a doctorate.[7]
The reunion was specific. Walden recalled that Yarbrough texted him around July 2017: “We have a huge opportunity to take this work and make it a huge part of how people develop in their careers.”[7] Yarbrough had already moved from counseling couples to coaching startup leaders. Problems with conflict, boundaries, ego, and communication appeared in both settings.
Torch did not begin as software. Yarbrough described its MVP as “a matchmaking service between coaches and employees via email.” The team paired people, sent them to Zoom, and invoiced monthly. That manual service produced recurring revenue and low churn before investors funded a platform around it.[8]
The founding insight was that coaching could reach below the executive suite if software lowered coordination costs and made progress legible to the employer paying the bill. “Coaching isn’t available to the people who are further down the ‘power stack’ within the organization,” Yarbrough said.[8] That tension between a private, high-trust relationship and an employer's demand for measurement shaped the product from the start.
Timeline
- 2017: Yarbrough and Walden founded Torch and tested coaching through an email-and-Zoom concierge service.[7]
- Winter 2018: Torch joined Y Combinator.[4]
- April 2019: Torch raised a $10 million Series A and reported 100 customers.[1]
- March 2020: Torch and Everwise combined under the Torch brand.[9]
- February 2021: A $25 million Series B brought financing to nearly $40 million.[2]
- May 2022: A $40 million Series C followed a year in which Torch said revenue and headcount more than doubled.[3]
- August 2024: Heather Conklin became CEO; Yarbrough became executive chair.[5]
- 2025–2026: Torch remained commercially visible despite YC's inactive label.[4][6]
What They Built
The first software product turned a coaching engagement into a measured program. Torch interviewed a participant's manager, peers, and reports, converted that feedback into development goals, and gave both participant and coach access to the goals. Follow-up surveys asked colleagues whether behavior had changed. An HR dashboard showed the buyer participation and progress without exposing the substance of private coaching sessions.[8]
The service component remained central. Participants met coaches by video, while the software handled matching, assessments, goals, and program reporting. In 2019 the company charged roughly $500 to $1,500 per participant per month, depending on coach seniority and region.[1]
The Everwise combination widened the product. Everwise brought mentoring, peer support, structured learning, and a network of more than 6,000 mentors. Torch could now sell a learning portfolio instead of one coaching format: executive coaching for senior leaders, mentoring for broader employee groups, and centralized administration for L&D teams.[9]
