Contents07 sections
Steering committee, the decision for the carve-out has essentially been made. Then the head of the business unit speaks up: “If this thing takes off, we've just built ourselves a competitor.” Legal nods: “Then we need control rights. Consent requirements, a board seat, a right of first refusal.” The corporate venture unit adds: “But we won't be doing any follow-on funding.”
After twenty minutes everyone wants control, and nobody wants to pay for it.
The question isn't whether the corporate stays connected to the venture. The question is how. And most people answer it the wrong way round.
Why it didn't work internally
Almost every technology that a carve-out gets negotiated over has a backstory. Usually it goes: “We tried it internally.” That's almost always true. And it isn't proof of a lack of ability, but proof of a structural problem.
Clayton Christensen and Joseph Bower described this in one of the most cited studies in management research. Energy and resources in a company flow to the programs that serve powerful existing customers. Projects for technologies that don't have customers yet wither for lack of momentum and resources. Nobody has to do anything wrong for that to happen. The system does it on its own.
The numbers fit. In a 2021 survey of 1,178 top executives, McKinsey reported that more than four out of five new businesses fail to reach $50 million in annual revenue even at least four years after launch. These are self-reported figures, not audited ones. The direction is clear nonetheless.
In fairness: Christensen and Bower also show that established companies can succeed with radical technology when their existing customers want it. Not everything new fails internally. What fails is the new thing that has no customer in-house. Those are exactly the candidates for a carve-out.
Separating alone isn't enough either
The obvious conclusion would be: get it out, clean cut, done. That doesn't work well either.
Charles O'Reilly and Michael Tushman studied 35 attempts to produce breakthrough innovations in established companies, spread across 15 business units in nine industries. More than 90 percent of the attempts in so-called ambidextrous structures succeeded: their own unit, their own processes, but connected to the company through top management. In functional structures it was 25 percent. And the two extremes? Cross-functional teams inside the old organization and completely independent teams without support both came in at zero.
The message is clear: separate, yes. Cut off, no.
The research on spin-offs: the middle beats the extremes
For spin-offs themselves, research shows the same pattern, from several directions.
Semadeni and Cannella studied 142 spin-offs from 1986 to 1997. Their finding: the spun-off companies benefit from some links to the parent, but too many links hurt performance. There's a balance between too much and too little interference.
Sapienza, Parhankangas and Autio reach the same conclusion via knowledge across 54 Finnish industrial spin-offs: growth is highest when the spin-off's knowledge base partially overlaps with the parent's. Too little overlap slows learning, too much prevents new combinations.
And then there's Henry Chesbrough's study of Xerox's technology spin-offs. The higher Xerox's initial stake, the worse the performance. Spin-offs that Xerox steered looked for their markets close to Xerox's business. Those steered by investors searched more broadly. And that's exactly what a technology needs that found no place in the core business: a market beyond the core business.
Supplier yes, main customer no
So what does the right connection look like? One study offers a surprising pointer. Emilie Feldman examined what happens when the same people sit on the boards of both parent and spin-off. On average, it helps the combined stock returns of both companies. But the higher the share of revenue the spin-off makes with the parent, the more this dual role hurts the spin-off. If you depend on the parent as a customer and have its people on your board at the same time, you become an extended workbench. And on terms that benefit the parent.
That leads to a precise rule. The corporate as supplier is a strong connection: it supplies materials, components or manufacturing know-how, earns from the venture's growth and has a real interest in seeing it grow. The corporate as dominant customer with a board seat is a dangerous connection. As one of several customers, it's welcome. As the only one, it makes the venture dependent.
The supplier role has another advantage, and it's political. It defuses the business unit's worry. The feared competitor becomes a buyer. Then there's knowledge: Agarwal and colleagues show that the parent company's capabilities at the time of founding increase a spin-out's chances of survival. Their data comes from start-ups founded by former employees in the hard disk drive industry, not from carve-outs. The logic carries over all the same: a supply relationship keeps exactly this connection alive without delivering control along with it.
Seven people, seven worries
That leaves the question of why so many carve-outs still end up with control rights. In my experience, it's because a carve-out is never decided by one person. In 2025, Gartner measured groups of five to 16 people from up to four functions for complex B2B decisions, and 74 percent of these groups showed unhealthy conflict. A carve-out is no different. And everyone at the table has a different worry:
- Head of R&D: “Will this pull my people and my budget away from the core business?”
- IP management: “Why don't we just license it out?” The answer is in License or Company.
- Corporate development and CVC: “Do we have to co-fund the rounds?” If the structure is right, no. More in Dead Equity.
- Business unit: “Are we giving away something we'll miss?” No. You lose the operational burden, not the asset. And you become a supplier.
- CFO: “What does it cost, and what does it do to our numbers?” See Burn, Not Balance Sheet.
- Legal: “Do we own this cleanly enough to hand it over?” The question that takes longest and is most often underestimated.
- Inventors: “What's in it for me, and how risky is it?” The most honest answer: a majority in your own company and a real risk.
The champion usually sits in R&D or IP management. But the decision is made by the CFO and corporate venture, and legal takes the longest. If you answer each of these questions before it's asked, you won't have to hand out control rights as a sedative at the end.
The order afterwards
Research and practice add up to a clear picture of the corporate's role after the carve-out. A minority stake that pays off but doesn't block any round. An exclusive license for a clearly cut field of use. A supply relationship that earns from the growth. An observer seat instead of a steering mandate. And an option on the next project from its own lab.
That's more connection than a sale and less than a subsidiary. That middle is exactly where the results are.
If you supply, you share in the success. If you control, you slow it down. Give the venture a supply chain, not a leash.
Sources
- Christensen, Bower, “Customer Power, Strategic Investment, and the Failure of Leading Firms”, Strategic Management Journal 17(3) (1996): resources flow to programs for powerful customers, projects without customers wither; established firms succeed with radical technology when existing customers want it — doi.org
- McKinsey & Company (Leap by McKinsey), “The State of New-Business Building” (2021), via press release: survey of 1,178 C-level executives, more than four out of five new businesses below $50 million in annual revenue at least four years after launch; self-reported — prnewswire.co.uk
- O'Reilly, Tushman, “The Ambidextrous Organization”, Harvard Business Review (April 2004): 35 breakthrough projects in 15 business units, over 90% success in ambidextrous structures, 25% functional, 0% for cross-functional teams and for independent teams without support — hbr.org
- Semadeni, Cannella, “Examining the performance effects of post spin-off links to parent firms: Should the apron strings be cut?”, Strategic Management Journal 32(10) (2011): 142 spin-offs 1986–1997, some links help, too many hurt — doi.org
- Sapienza, Parhankangas, Autio, “Knowledge relatedness and post-spin-off growth”, Journal of Business Venturing 19(6) (2004): 54 Finnish spin-offs, growth highest with partial knowledge overlap — doi.org
- Chesbrough, “The governance and performance of Xerox's technology spin-off companies”, Research Policy 32(3) (2003): 35 spin-offs, a higher initial Xerox stake goes along with worse performance, Xerox-steered spin-offs look for markets close to the core business — doi.org
- Feldman, “Dual directors and the governance of corporate spinoffs”, Academy of Management Journal 59(5) (2016): dual directorships positive on average for combined returns, but increasingly negative for the spin-off the higher its share of revenue with the parent — doi.org
- Agarwal, Echambadi, Franco, Sarkar, “Knowledge Transfer through Inheritance: Spin-out Generation, Development, and Survival”, Academy of Management Journal 47(4) (2004): parent company capabilities increase the knowledge base and survival chances of employee spin-outs (hard disk drive industry 1977–1997) — doi.org
- Gartner, “Gartner Sales Survey Finds 74% of B2B Buyer Teams Demonstrate Unhealthy Conflict During the Decision Process” (press release, May 7, 2025): buying groups of five to 16 people from up to four functions, 632 B2B buyers surveyed — gartner.com
