How should large companies build new businesses alongside a dominant core?

Jag Duggal · Mar 29, 2023

New product teams should be deliberately kept small and told to build something fundamentally different rather than incrementally better, with further headcount unlocked only after customer love is demonstrated

Incrementally better won't get anyone to notice, and intentionally limiting resources to a small group of hyper-driven people forces creativity and preserves autonomy

Scope: teams of roughly a dozen at launch

37:22 20Product: Nubank's CPO on Why Product is 90% Science and 10% Art, Why Execution is Overrated and Strategic Clarity is Under-Appreciated, Why You Should Never Fall in Love With Your Own Ideas & Nubank's Biggest Product Challenges Scaling to 80M Users

Andrew Dudum · Apr 4, 2026

Hims & Hers should be run as a public shell for innovation — a portfolio of increasingly independent bets that are starved, funded, or ring-fenced like a venture incubator

What constitutes great health and wellness is constantly changing and accelerating — new diagnostics, devices, drugs, protocols — so the company must be a constantly evolving portfolio to curate the best and deliver it affordably at scale

13:51 20VC: Hims & Hers: $4.3BN Market Cap on $2.3BN of Revenue: The Comeback | Why Being Public is 10x Better | The Death of the "Strategy" Hire | Why Performance Marketing is Worse than Brand Marketing with Andrew Dudum

Andrew Macdonald · Aug 17, 2026

New businesses inside large companies should operate on weekly cadences and be forced to re-ask for funding, because abundant resources make them slower and more cash-consuming than first-principles startups

People get fat on the resources of a big balance sheet, so they don't build it the way a startup would; they end up slower and constantly chasing the people building from first principles

14:31 20VC: Uber President on The Untold Uber Stories: Travis, China and Self-Driving | Why Autonomy Is Existential | How to Beat DoorDash to #1 in Food with Andrew MacDonald

Zach Perret · Oct 16, 2024

New product efforts should be staffed as the smallest sufficient 'atomic team' and deliberately hidden from the rest of the company until they reach a first design partner

If the whole company learns about a new team, hundreds of well-meaning 'how can I help' messages effectively DDoS the team and stop it from moving fast

Scope: founder must still supply sufficient resources and protection

9:09 20VC: Why Founder Mode is Dangerous & Could Encourage Bad Behaviour | Why Fundraising is a Waste of Time & OKRs are BS | Why Angel Investing is Bad for Founders to Do and the VC Model is on it's Last Legs with Zach Perret @ Plaid

Matthew Steckman · Mar 23, 2026

The right way to attack a new defense market is a small tiger team funded by internal R&D that builds a rough demonstrator quickly — 'a dart to throw' — then pulses the market to see how close it landed, iterating or scrapping from there.

You match your own internal motivation to what the market is whispering, and only learn what the customer actually cares about by putting something concrete in front of them.

Scope: framed as how Anduril does it

24:57 20VC: Inside Anduril's $20BN Army Contract & Why Anduril Must Go Public | Why 99% of Drone Companies Will Die | Why There is Never an Ethical Question of How Anduril Products are Used with Matthew Steckman, President @ Anduril

Jackie Reses · Mar 31, 2023

Old-school infrastructure companies and large financial services firms will be the losers unless they set up teams to disrupt their own products; if they instead manage for cash flow they will die a long-term death

They understand they need to innovate, but the open question is whether they can actually execute self-disruption

Scope: they do have a chance if willing to disrupt themselves

32:56 20VC: The Secret to Negotiating; Making $50BN for Yahoo on Alibaba | Why Everything You Know About Hiring is Wrong; Domain Knowledge and Past Experience are Dangerous | The Next 10 Years for Fintech; Winners, Losers and Crypto with Jackie Reses, CEO @ Lea

Martin Gontovnikas · Feb 14, 2024

Companies never get too big to take big bets — the large companies that keep succeeding are the ones still taking them, and the rest eventually get killed by a startup

Amazon's willingness to ship a failure like the Kindle phone is what made it willing to try AWS, now its main profit engine

15:27 20Growth: How to Master Product-Led-Growth, The Biggest Mistakes Startups Make When Scaling into Enterprise, How to Assess "Bets" in Growth; Which to Take and Which to Not with Gonto, Interim CMO @ Vercel

Dharmesh Shah · Jun 13, 2022

Every organization, team and individual must allocate a non-zero percentage of resources to outlier bets with disproportionate returns — the amount is negotiable, zero is not

Good managers will always run the existing predictable business well (Innovator's Dilemma), so without a deliberate allocation you get caught by regression to the mean and end up wishing you had dug the well before the house was on fire

Scope: 80-95% of resources on predictable high-return work is fine; the battle must be won at the macro/organizational level first; leaders shouldn't dictate which bets, only that the percentage isn't zero

41:47 20VC: Hubspot Co-Founder Dharmesh Shah on The 3 Risks All Startups Face, Angel Investing Rules; No Founder Meetings and No Due Diligence, SMB vs Enterprise; Lessons on Pricing, Distribution and Why You Should Resist Going Enterprise

Andrew Macdonald · Aug 17, 2026

Incubating new businesses requires dedicated headcount (roughly 5-7% of a business unit's people) rather than asking existing leaders to do it part-time

Trying to incubate something with 2% of your time while running a core marketplace is really hard; it requires dedicated capacity and thinking

Scope: numbers given as illustrative

13:37 20VC: Uber President on The Untold Uber Stories: Travis, China and Self-Driving | Why Autonomy Is Existential | How to Beat DoorDash to #1 in Food with Andrew MacDonald

Geoff Charles · Dec 6, 2023

It is very hard for large companies to actually innovate.

0:00 20VC: Ramp's Product Playbook: How To Hire Product Teams, How to Run Sprints, How to Increase Product Velocity, When and How to Go Multi-Product with Geoff Charles, VP Product @ Ramp

Geoff Charles · Dec 6, 2023

Large incumbents find it extremely hard to innovate or launch new products, even with huge resources

They face short-term board expectations, a go-to-market team incentivized around one business model, understanding of only one customer profile, and an engineering DNA built around existing orgs — launching something new requires breaking all of those at once

34:46 20VC: Ramp's Product Playbook: How To Hire Product Teams, How to Run Sprints, How to Increase Product Velocity, When and How to Go Multi-Product with Geoff Charles, VP Product @ Ramp

Sanjit Biswas · Dec 8, 2023

Companies should maintain a 'ladder of risk' — funding a couple of high-risk, high-reward new product bets alongside the low-risk core — even when the core product still has bugs and unmet feature asks

Building for the long term means planting seeds you harvest years out; the prototype and V1/V2 cost little money but take time, and Samsara's second product (safety) is now its number one revenue product

33:08 Ladder of risk fund long horizon bets before the core is polished

Zach Perret · Oct 16, 2024

Running internal new-product funding as a formal VC-style milestone process doesn't work; it's better to let teams run at the problem and come back when they know

Over-processizing makes everyone hack the process instead of finding the unique insight, and it forces teams onto artificial timelines when a real customer may be ready in a week

10:23 Avoid formal milestone funding let teams self report when ready

Jackie Reses · Mar 31, 2023

Incumbents won't build these products because their embedded infrastructure and the benefits they derive from the existing system mean they choose not to disrupt themselves

Same reason JPMorgan didn't create Square

32:27 Incumbents rationally choose not to disrupt themselves due to embedded infrastructure benefits

Jackie Reses · Mar 31, 2023

PayPal, Square, Adyen and Stripe are now incumbents but are the ones doing self-innovation best, building multibillion-dollar products inside their own organic infrastructure

There is a magic to how they generate new products at scale that others should learn from

33:26 Leading payment companies paypal square adyen stripe exemplify successful self disruption

Jackie Reses · Mar 31, 2023

A culture of disruption requires making individual contributors as celebrated as managers and rewarding people for leaving big jobs to start something small

The convention that power grows with headcount underneath you is a cancer in a company and a culture killer on invention

34:38 Celebrating individual contributors not just managers sustains a culture of disruption

Adam Mosseri · Jul 24, 2023

Threads is still a genuine risk, and its opportunity cost is larger than the small team size suggests.

Failure can damage reputation and brand and make people less likely to want you to succeed later; and although the team is only dozens of people, they are extremely good engineers, designers, PMs, data scientists and researchers, so headcount understates the cost.

51:01 Opportunity cost of talented small teams on new bets is larger than team size suggests

Daniel Khachab · Oct 28, 2024

Special projects / skunkworks teams are the wrong way to build a company; innovation must happen in the core

If the core lacks the capability to innovate you already have a problem, and a skunkworks team is an excuse for that; also the person who figures something out should own and scale it, which is far more motivating than handing it to others to productize

52:19 Skunkworks teams are wrong innovation must happen in the core

Matthew Steckman · Mar 23, 2026

Conviction to concentrate capital should come from an accumulation of proof points — formal and informal customer feedback, operator feedback, and successful engagement with Congress — rather than any single signal

In US defense the funder is Congress, so legislative buy-in is part of the evidence base alongside customer and operator signals

26:58 Concentrate capital only on accumulated customer operator and legislative proof points

Matthew Steckman · Mar 23, 2026

A moderately sized system like Roadrunner costs well over $100M before generating any revenue, so the right discipline is to not enter the development curve at all unless you believe the product will come out the other side.

Once you're on the curve you're $10M, $20M, $50M, $100M+ in, so Anduril kills developmental programs well before they reach high spend rather than cutting them late.

43:25 Kill development programs before spend ramps not after

Aman Narang · Aug 21, 2024

Running separate teams that scale the existing business and teams that tinker on new bets is what lets a company survive many failed product attempts

33:03 Separate scaling and new bet teams enables surviving many failed attempts

Aman Narang · Aug 21, 2024

For an internal new-ventures program to work, the benefits of being inside the parent company must outweigh the overhead of being part of a large organization, so new ventures should be kept as separate from the core business as possible

People who would otherwise leave to start VC-backed companies only stay if the upside of the platform beats the drag of big-company overhead

33:18 Internal venture benefits must outweigh corporate overhead so keep ventures separate from core

Andrew Macdonald · Aug 17, 2026

Uber's scale means any new product must have a path to multiple billions of dollars of GMV within a few years to be worth pursuing, which constrains the company's thinking and risks preventing it from trying new things.

At near-$1 (hundred-billion-scale) gross bookings, nothing smaller passes the significance test

12:05 Scale hurdle rate screens out small new bets

Andrew Macdonald · Aug 17, 2026

Large companies struggle to build new businesses because the existing core swallows organizational capacity, resourcing, distribution and management focus

The existing business is so big it consumes engineering, marketing dollars and management attention; hard to focus on the new thing while managing a $225 billion business

12:54 Core business swallows capacity and management focus

Harry Stebbings · Aug 17, 2026

Nick Storonsky is the best founder in the world at running new-product experiments, and his method of funding 26 parallel experiments with $2M each and weekly 20-minute check-ins is the model

Having interviewed a thousand founders, he is the single best one interviewed

14:11 Many parallel small bets with weekly check ins