首页  >>  来自播客: Sequoia Capital 更新   反馈  

Sequoia Capital - Databricks’ Ali Ghodsi Never Wanted to Be CEO. Now He’s Among the Best

发布时间:   原节目
以下是原文的中文翻译: 在最近一集《Long Strange Trip》节目中,Databricks的创始人兼首席执行官Ali Goazi分享了他从教授到领导一家数十亿美元公司的历程见解,为首次担任首席执行官的人提供了宝贵的经验教训。a16z的Ben Horowitz曾称赞Goazi是“最优秀的首席执行官”,这证明了他的变革性领导力。 Goazi成为首席执行官的道路不同寻常。2015年,Databricks尽管凭借Apache Spark取得了开源成功,但在商业上却举步维艰,差距收入仅为150万美元。董事会寻找新的首席执行官,面试了外部候选人,而联合创始人Goazi则考虑重返学术界。他将自己接任首席执行官的决定归因于一种人生模式:选择挑战性的道路而非舒适的道路,最终拥抱商业世界的未知。 执掌公司后,Goazi发现了一个关键瓶颈:商业上的不成功。他做出了一个关键决定,在18个月内彻底改组了整个高管团队,意识到之前的“产品主导增长”(PLG)战略不足以应对。他强调招聘企业销售专业人才,此举与公司最初反销售的理念相悖。这使他找到了Ron Gabisco,一位兼具激进销售敏锐性和工程背景的销售主管。Goazi了解到,成功的销售人员拥有“专业进取心”、高情商以及理解公司“权力基础”的能力——这是一种与技术专长不同的艺术。他的高管招聘策略侧重于极度挑剔,提前(6-12个月)开始物色人选,并进行广泛的“暗中”背景调查,以避免代价高昂的“假阳性”。 Databricks与Snowflake的竞争是另一个决定性时期。Goazi描述了一个细致的策略:“仔细研究你的对手,了解他们的弱点,并将你的优势应用于他们的弱点。” Snowflake的专有锁定、糟糕的AI支持和高成本被认定为弱点。Databricks则以“开放湖仓”愿景反击,该愿景自2009年公司成立之初就植根于AI,并提供显著更低的总拥有成本(TCO)。这一策略促成了“湖仓”类别的创建,这个术语最初遭到嘲笑,但由于Databricks“狂热般的执着”在全公司范围内的关注,最终被业界广泛接受,即使这意味着牺牲短期的广告投资回报率。Goazi强调了创新而非简单复制的重要性,确保他们的产品真正与众不同,而不仅仅是竞争对手产品的“更好”版本。 关于领导风格,Goazi承认自己有“杀手”的声誉,并将其归因于他的移民背景带来的一种不甘示弱的心态。他强调,回避冲突是“首席执行官的癌症”,主张进行清晰透明的沟通,以防止组织漂移。他认为管理冲突类似于锻炼或健康饮食——困难但必要。 谈及规模化发展,Goazi指出,当公司员工达到约250人时,就会从“明星”模式(首席执行官了解一切)转变为需要健全的流程和“经理的经理”。他还谈到了现代组织设计趋势,例如Jack Dorsey所探讨的那些。虽然他同意AI作为组织智能的概念(Databricks的“Genie”产品构建了一个“本体”来为AI提供企业上下文),但他对过于剧烈地削减人类组织层级表示怀疑,特别是在经理监督25名直接下属的同时还扮演“球员兼教练”的角色,他引用了管理中不可或缺的人性因素。 Goazi继续编码,甚至承担生产级别的工作,以亲身了解工程瓶颈,他认为重新设计组织流程对于AI的采用至关重要。他的日常工作优先处理“主要瓶颈”,而非被动地管理日程。尽管拥有巨大的私人估值,Databricks仍保持私有化,Goazi解释说,他宁愿在远离公众审视的情况下应对当前“疯狂的AI转型”和市场波动,他认为公众很难理解这种革命性的转变。

In a recent episode of "Long Strange Trip," Ali Goazi, the founder and CEO of Databricks, shared insights into his journey from professor to leading a multi-billion dollar company, offering valuable lessons for first-time CEOs. Ben Horowitz of a16z famously touted Goazi as the "best CEO out there," a testament to his transformative leadership. Goazi’s path to CEO was unconventional. In 2015, Databricks, despite open-source success with Apache Spark, struggled commercially, with only $1.5 million in gap revenue. The board sought a new CEO, interviewing external candidates, while Goazi, a co-founder, contemplated returning to academia. He credits his decision to take on the CEO role to a life pattern of choosing challenging paths over comfortable ones, ultimately embracing the unknown of the business world. Upon taking the helm, Goazi identified a critical bottleneck: the lack of commercial success. He made a pivotal decision to revamp the entire executive staff within 18 months, realizing the previous "product-led growth" (PLG) strategy was insufficient. He emphasized hiring enterprise sales professionals, a move that contradicted the company’s initial ethos against sales. This led him to Ron Gabisco, a sales leader with a rare blend of aggressive sales acumen and an engineering background. Goazi learned that successful salespeople possess "professional aggression," high emotional intelligence, and the ability to understand a company's "power base" – an art distinct from technical expertise. His executive hiring strategy focused on extreme pickiness, starting searches early (6-12 months), and conducting extensive "backdoor" reference checks to avoid costly "false positives." Databricks' rivalry with Snowflake was another defining period. Goazi described a meticulous strategy: "study your enemy carefully, understand their weaknesses, and apply your strength to their weaknesses." Snowflake’s proprietary lock-in, poor AI support, and high cost were identified as vulnerabilities. Databricks countered with an "open lake house" vision, rooted in AI from their inception in 2009, and a significantly lower Total Cost of Ownership (TCO). This strategy led to the creation of the "Lakehouse" category, a term initially ridiculed but eventually embraced industry-wide due to Databricks' "maniacally religious" company-wide focus, even if it meant sacrificing short-term ad ROI. Goazi highlighted the importance of innovation over mere copying, ensuring their offering was truly different rather than just a "better" version of a competitor's product. Regarding leadership style, Goazi acknowledged a "killer" reputation, attributing it to a "chip on his shoulder" from his immigrant background. He stressed that conflict aversion is "cancer in a CEO," advocating for crystal clear communication to prevent organizational drift. He views managing conflict as akin to exercising or healthy eating – difficult but necessary. On scaling, Goazi noted that around 250 employees, a company shifts from a "star" model (where the CEO knows everything) to needing robust processes and "managers of managers." He also touched on modern organizational design trends, like those explored by Jack Dorsey. While agreeing with the concept of AI as an organizational intelligence (Databricks' "Genie" product builds an "ontology" to feed AI enterprise context), he expressed skepticism about collapsing human organizational layers too drastically, especially regarding managers overseeing 25 direct reports while also being "player-coaches," citing the essential human element of management. Goazi continues to code, even committing production-level work, to understand engineering bottlenecks firsthand, observing that re-engineering organizational processes is critical for AI adoption. His daily routine prioritizes the "main bottleneck" over reactive calendar management. Despite immense private valuation, Databricks has remained private, with Goazi explaining that he prefers to navigate the current "crazy AI transformation" and market volatility away from public scrutiny, which he believes struggles to comprehend revolutionary shifts.