The Serial Builder Advantage: Why Repeat Innovators Win

[McKinsey & Company video description] In an era of economic uncertainty and accelerating technological disruption, creating new businesses has become less of a growth option and more of a strategic imperative. In this episode of The McKinsey Podcast, McKinsey Senior Partner Jason Bello speaks with Editorial Director Roberta Fusaro about McKinsey’s new research on corporate venture building. They examine why serial builders outperform their peers, how AI is reshaping the economics of innovation, and what leaders can do to turn experimentation into a repeatable capability.

TL;DR

A ~25-minute McKinsey Podcast interview: Jason Bello (McKinsey Senior Partner) in conversation with Roberta Fusaro (Editorial Director) on new McKinsey research into corporate venture building — companies launching wholly new, step-out businesses rather than incremental product extensions. Eight load-bearing claims:

1. Serial builders dramatically outperform one-shot builders. Companies that build three or more ventures at once “dramatically outperform” those that only try once — Bello’s soccer analogy: “multiple shots on goal against a single goal” beats either scattering bets everywhere or betting everything on one shot.
2. Break-even cost is collapsing fast. The average cost to reach break-even on a typical new venture fell from ~$125M (2024) to ~$77M (2025) — driven by (a) capital-constrained organizations demanding earlier proof of value, and (b) AI.
3. Two flavors of AI in venture building. “AI as a business” — building a fully agentic AI-native version of yourself, deliberately structured as a new venture even when it does the same work as the core business (Bello’s example: a bank building a fully-agentic B2B bank — KYC, customer service, and flow of funds all automated end-to-end — with the intent to eventually reintegrate it). “AI in the background” — co-pilots for every step of building: McKinsey’s own Beacon platform for product management and acceptance-criteria generation, AI-assisted business-case pressure-testing, marketing/AB-testing content, and fractional AI-played roles (e.g. a sales-ops function) before a full-time hire is justified.
4. Ideation quality, not just speed. AI broadens the “aperture” during ideation — surfacing analogous moves from other industries/customer segments — and enables synthetic customer personas built from real public data (Bello’s pet-care example) to pressure-test ideas before real user research. Caveat, stated explicitly: synthetic personas tend toward sycophancy (“they love everything”) and are a supplement to, never a replacement for, real user research.
5. Prototyping has compressed from weeks to a day. Where wireframes once wowed clients over an overnight turnaround (as recently as ~24 months prior), teams now arrive with a vibe-coded, working prototype built in a day — both a better experiential anchor and something that can be market-tested directly.
6. Milestone-based, tranche funding replaces annual budgeting. Ventures are funded by dividing a total investment horizon (e.g. 9 months) into chunks (e.g. three 3-month tranches), checking in on whether milestones were met, and releasing the next tranche accordingly — de-risking and enabling faster kill/pivot decisions.
7. A fact-based, blame-free culture is the enabling condition. “If the facts tell us our product stinks, so be it” — attributing pivots to what was learned, not to personal failing, is what lets teams change course quickly without shame attached.
8. Disruption itself is the argument for venture building, not against it. In a period of uncertainty, failing to build new businesses raises the odds of being disrupted by someone else who does. Corporate builders also have structural advantages over independent startup founders: less time spent fundraising (the leadership team and board already know you) and ready-made incumbent assets (existing customer base, untapped proprietary data/IP).

What was actually ingested

The full ~25:30 episode via its auto-generated (ASR) English caption track — 614 transcript segments, no chapters provided by YouTube. Acquire-phase note: the initial headless fetch (both at the default 30s timeout and a retried 60s timeout) failed with transcript panel did not render — diagnosis via a direct network-response check confirmed YouTube’s youtubei/v1/get_transcript endpoint returned HTTP 400 failedPrecondition, the same class of automation-detection rejection documented in the skill’s 2026-05-13 incident note, despite the skill’s navigator.webdriver masking already being in place. A --headed retry (real Chromium window, this being a macOS host with a display) succeeded cleanly and recovered all 614 segments. Light ASR cleanup applied to proper nouns: “McKenzie” → “McKinsey”, “Jason Bellow” → “Jason Bello”, “Robera Fisaro” → “Roberta Fusaro”, plus a handful of concatenated/misheard terms (ideiation → ideation, sick ofantic → sycophantic, vibecoded → vibe-coded, milestonebased → milestone-based).

Why this source matters to the wiki

This is the wiki’s sharpest quantified case for the digital-seizing/balancing-digital-portfolios microfoundation in the dynamic-capabilities / warner-wager-process-model framework — Bello’s research is a cross-firm, multi-company empirical claim (not a single case study) that directly measures the payoff of a portfolio approach to venture building, and pairs it with a concrete, falling cost-to-break-even trend. It also supplies:

• A crisp two-way taxonomy of AI-in-venture-building (“AI as a business” vs. “AI in the background”) that is more granular than the wiki’s existing enterprise-AI-adoption material’s typical build/buy framing — see enterprise-ai-adoption.
• A concrete, named mechanism (milestone-tranche funding, ring-fenced reporting line to a C-suite exec) for digital-seizing/rapid-prototyping and strategic-agility, directly comparable to DBS’s Innovation Pyramid + QPR/slush-fund system.
• A sharply-worded articulation of the fact-based, blame-free culture mechanism for strategic-renewal/organizational-culture — “get the facts on the table… there’s no fingerpointing” — that sits alongside Carroll’s academic culture-as-social-control-system theory and Erginbilgiç’s non-digital performance-culture case.

Linked entities and concepts

• dynamic-capabilities — sharpest quantified case yet for the balancing-digital-portfolios microfoundation.
• warner-wager-process-model — tagging vocabulary source; see dynamic_capabilities: above.
• enterprise-ai-adoption — the AI-as-a-business / AI-in-the-background taxonomy is a venture-building-specific companion to the broader enterprise-adoption material.
• McKinsey & Company — publishing entity.
• Dangling (single-source mention, deferred per Author-entity promotion): Jason Bello (McKinsey Senior Partner, interviewee) — first wiki mention; promote on second-source mention. Lucia Rahilly (named in the stock sign-off outro only, not an active interviewer in this episode) — first wiki mention; deferred.
• Roberta Fusaro — Editorial Director; interviewer/host for this episode. Promoted to an entity page on this ingest — her second appearance across McKinsey Podcast sources in this wiki (first: credited in body prose on 2026-06-18-ramaswamy-mckinsey-every-company-software-company; here she is the named, active on-camera interviewer).

Source quality

Auto-generated (ASR) captions, standard fidelity for the wiki’s video corpus — recovered only via a --headed retry after headless fetches were rejected by YouTube’s transcript API (see Acquire-phase note above). Content is a first-party McKinsey Podcast interview built around the firm’s own proprietary research (unpublished/undetailed methodology for the $125M→$77M and “3+ ventures” findings — no sample size or survey instrument named in the interview itself); treat the headline numbers as directionally credible (consistent with McKinsey’s broader track record in this wiki, e.g. 2026-04-28-brynjolfsson-canaries-coal-mine’s independent corroboration pattern) but not independently verified — no linked underlying report was found at ingest time. No sponsorship beyond McKinsey’s own platform.