This CPO regrets that product management exists | Tom Verrilli (CPO of Whatnot)

The podcast discusses the evolving role of product management, arguing that while PMs can provide valuable leverage in complex areas, their overuse can hinder engineers and designers by depriving them of decision-making opportunities. The idea that "we regret product management exists" is presented not as a rejection of the role, but as a mindset to avoid unnecessary PM hires - favoring lean, context-driven teams where PMs are assigned only when and where needed. There's a growing shift toward fewer, more senior PMs who act as individual contributors, focusing on high-impact decisions rather than process and alignment, with AI tools enabling faster execution and reducing reliance on traditional oversight.

The discussion emphasizes empowering engineers and designers to take ownership of product decisions, advocating for flexible team structures where responsibilities are shared across roles. Companies like What Not serve as case studies in minimizing PM headcount while maintaining rapid innovation through real-time adjustments and direct customer engagement. The conversation also explores mental models for strategic thinking - such as "playing the accordion" between planning and experimentation - and stresses the importance of hands-on leadership, deep system understanding, and avoiding decisions based solely on averages. Ultimately, the focus is on building adaptable, high-leverage teams where product thinking is distributed, and value comes from impact, not role specialization.

• What if you led a high-impact experiment without a PM?

  • Move: Identify a small but strategic product area (e.g., onboarding flow or notification logic) where you can define the problem, design a solution, run an A/B test, and analyze results - all solo using AI tools for data pulls and user behavior insights.
  • Why Now?: AI now enables solo operators to do what once required PMs, data scientists, and analysts - pull cohorts, build models, and validate assumptions in minutes, not weeks.
  • Expected Upside: You ship faster, gain full ownership of the decision loop, and compound learning - building both product judgment and execution speed while proving that PM-like outcomes don't require a PM.

• What if you rebuilt your next feature live, with real-time feedback?

  • Move: Instead of writing specs and waiting for reviews, launch a minimal version of your feature behind a config flag, then observe real user interactions in real time using AI-powered session analysis tools; adjust logic or UI within hours based on actual behavior.
  • Why Now?: Platforms like Whatnot show that live product tuning (e.g., adjusting auction timers mid-stream) is not only possible but optimal - AI now gives solo developers instant access to behavioral data and debugging context.
  • Expected Upside: You cut cycle time by 50 - 80%, eliminate guesswork, and build products that respond to real needs - not assumptions - creating tighter feedback loops and higher user satisfaction.

• What if you became the DRI (Directly Responsible Individual) for a compounding business lever?

  • Move: Pick one high-leverage metric (e.g., conversion rate, retention, or GMV per session) and own it end-to-end: define the hypothesis, build the solution, run experiments, and report impact - without delegating any part of the chain.
  • Why Now?: The trend toward fewer, more senior PMs and IC-heavy teams means top performers are expected to drive outcomes, not just output. AI tools let you simulate downstream effects and assess risks mentally before shipping.
  • Expected Upside: You create outsized value relative to team size, demonstrate cross-functional mastery, and position yourself as a force multiplier - exactly the profile of the ICs shaping the future of product-led companies.

• Adopt a "no PM by default" mindset and only take on product management tasks when a clear, specific problem requires coordination, ownership, or decision-making that can't be handled by engineering or design alone.
• Develop systems thinking by practicing mental modeling: before building, ask what happens if the feature succeeds or fails, how it scales 100x, and what downstream risks (legal, financial, UX) might emerge - then act decisively.
• Use AI tools like Claude or Hex to independently analyze user data, understand codebases, and test product flows without blocking on engineers or data scientists, reducing dependency and speeding up iteration.
• Focus on high-leverage, compounding experiments rather than incremental improvements; prioritize projects where success unlocks multiple follow-up opportunities and long-term business impact.
• Stay hands-on by regularly engaging with real user feedback, support tickets, and live product behavior - treat firsthand experience as non-negotiable for making accurate product decisions, especially in fast-moving environments.