The McKinsey Podcast

The episode contains a handful of useful data points and framing (cost-to-break-even drop, portfolio mindset, AI-native build-to-disrupt-yourself) but is padded with well-worn advice about culture, psychological safety, and milestone tracking that any experienced operator will have heard many times. The ratio of genuinely new claims to filler is modest for a 26-minute runtime.

“In 2024, the average cost to reach break even on the typical new venture was around $125 million. In 2025, that number dropped to $77 million.”

“players who take multiple shots on goal against a single goal tend to be much more successful than players that are placing lots of bets all over the place”

Most frameworks presented - portfolio mindset, milestone-based funding, 'building the muscle,' psychological safety - are well-established consulting tropes. The 'build an AI-native version of yourself to disrupt yourself' angle has some freshness, and the synthetic customer persona experiment is a mildly novel operational detail, but nothing is genuinely contrarian or first-principles.

“build the business. You're actually not building a step out business in that case necessarily. You're building something that replaces you, but you're doing it like an attacker”

“we were able to set up synthetic pet owners because there's so much data out there about pet owners”

Jason Bellow is a senior McKinsey partner with clear hands-on advisory experience across healthcare and fintech venture builds, and he references specific client engagements credibly. However, he is primarily a consultant rather than an operator who has personally scaled a business, so his authority is advisory rather than direct practitioner experience at scale.

“I spent all my life in healthcare. I'm doing the same kind of things with healthcare payers and healthcare intermediaries”

“I've done a business where we built a new durable medical equipment device that helps with mobility, and I've seen people using it”

The episode has some genuine specifics - the $125M-to-$77M cost shift, the 'three or more ventures' finding, and the agentic KYC bank example - but all client cases are fully anonymised with no company names, no revenue outcomes, and no details on sample size or methodology for the underlying survey. Numbers appear but without enough context to fully evaluate them.

“In 2024, the average cost to reach break even on the typical new venture was around $125 million. In 2025, that number dropped to $77 million.”

“we found out that the market was an order of magnitude smaller than we thought it was”

The host moves the conversation along adequately and asks reasonable follow-on questions, but there is essentially no pushback, no challenge to any claim, and no probing on survey methodology, sample size, or the self-serving nature of McKinsey promoting its own Beacon platform. The tone is promotional throughout, and every guest statement is greeted with affirmation.

“How do you sort of get your personas to be real with you?”

“Does it still make sense to think about new business building in this market?”