Contents07 sections
A corporate decides to build a new business outside its own walls. Its own company, its own team, funding in stages, an option to buy it back later. Everything is set up properly. On the corporate side, a manager gets the topic on top of their day job. Every three months they sit on the board, read the slides, ask two good questions and fly home.
Two years later the venture has customers, revenue and a team. Time to bring it home. And suddenly nobody wants it. The business unit it was meant for has a new head and different priorities. The product overlaps with something another division has built in the meantime. The manager from the board has changed jobs.
The business was ready. The company wasn't.
The example is made up, the pattern isn't. And it has a clear cause: the venture was separated, but nobody was tied in.
Separation is right, isolation isn't
Let's be clear first: building a new business outside the core organisation is the right call. Charles O'Reilly and Michael Tushman looked at 35 breakthrough initiatives in large companies and compared four ways of organising them. Of the initiatives that sat in cross-functional teams or ran without real support, not one reached its goals. In classic functional structures it was 25 percent.
The winner was a design they called ambidextrous: more than 90 percent reached their goals. And the decisive part is in the definition: organisational separation through a tightly integrated senior team. Separate in the day-to-day, tied in at the top. Seven of eight initiatives that switched to this design improved substantially afterwards.
The sample is small, 35 cases. But the direction is clear. Separation alone isn't the recipe. Separation with a tie-in is.
Venture units die on missing relationships
Christopher Hill and Julian Birkinshaw followed 95 corporate venture units from 2001 onwards. Within roughly two years, 23 percent had been closed. What distinguished the survivors: active relationships with the parent's executives and with its business units. Not just with investors outside, but with the people inside who would one day have to take the business on.
Vijay Govindarajan put the same thing in one sentence: "There has to be a link." The innovation team should not be isolated. And then the honest second half: make the link, expect the conflict, and manage it.
A board seat is oversight, not ownership
Board seats are the norm. In a Global Corporate Venturing poll, 90 percent of corporate investors said they hold some kind of board position in their portfolio companies. That makes sense for a minority stake. You watch your investment, you protect your rights, you don't interfere.
But a business that's meant to come home is not a minority stake. It needs someone on your side who wants it. Not someone who watches it once a quarter. In my view, that's exactly the difference between a portfolio holding and your future business. And no study measures it as neatly as that, I'll admit. But the evidence above points in the same direction: whoever is tied in only at board level isn't tied in.
Even the consultancies say so, though here I'd take the numbers with a pinch of salt. In a McKinsey survey of 2023, the companies that build new businesses most successfully have at least one C-suite sponsor and hold regular meetings between the new business's leaders and the core organisation's senior leaders. That's a self-reported survey by a firm selling exactly this service, and the reported factor of 2.4 applies to a whole bundle of measures, not to sponsorship alone. It's a hint, not proof.
The gym, and why paying alone isn't enough
There's an image I like for this: a gym. Nobody gets fit watching through the window. And a membership that costs nothing is the easiest one to skip.
Now, fair is fair, because research makes the image more uncomfortable than it first looks. Stefano DellaVigna and Ulrike Malmendier analysed 7,752 gym members over three years. Those with a monthly flat rate of more than 70 dollars went on average 4.3 times a month, paying more than 17 dollars per visit, when a ten-visit pass would have cost them 10. They paid in advance and still didn't go. Paying alone doesn't make anyone train.
Ashraf, Berry and Shapiro found something similar in a field experiment in Zambia: a higher price doesn't make people use a product more. It filters. Those who pay more are those who wanted it anyway.
That doesn't refute the image. It sharpens it. A one-off payment is easy to forget. What works is twofold. First: money that's decided anew at every stage, gate by gate, tied to visible progress. Each tranche is a fresh decision that someone on the corporate side has to stand behind. That's the filter Ashraf describes, applied over and over. Second: you train yourself. If your own people don't work with the venture every week, they'll never feel what's growing there.
Three roles, named before the build
What does that mean in practice? Ownership on the corporate side has to be designed, not assumed. At the latest when an idea has been validated with real buyers, three roles need names:
The owning business unit. Who benefits once the venture exists? Whose numbers does it move? That's who it comes home to, so that unit has to want it from the start.
The board members. Who takes the strategic decisions, funds the next stage or stops it? Not as a side job, but as someone accountable.
The daily counterpart team. Who works with the venture week by week? Market, product, sales, digital. In my experience, four to six hours a week per person is enough, as long as it's the right people.
And a fourth role that's easy to forget: the connector. Your people know what's happening elsewhere in the group. The venture team never will. Without that link, the same business gets built twice in two parts of one company. And paid for twice.
Expect the conflict
One thing shouldn't be glossed over: tie-in creates friction. Cisco's spin-ins are the best-known example. The model worked commercially, but inside the company it caused resentment, and according to press reports the founders of the last spin-in left a few years after the buy-back. Proximity to the parent isn't free.
That's exactly why the roles need to be named early and the rules fixed up front. A conflict you've anticipated and regulated is manageable. A conflict that only surfaces on the day of the buy-back costs you the business.
If you only show up for the board meeting, the business never comes home.
Sources
- O'Reilly, Tushman, "The Ambidextrous Organization", Harvard Business Review 82(4), 74–81 (2004): 35 breakthrough initiatives in four designs; more than 90 % of ambidextrous designs (organisational separation through a tightly integrated senior team) reached their goals, none of the cross-functional or unsupported teams, 25 % of functional designs; 7 of 8 initiatives improved after switching to the ambidextrous design — hbr.org
- Hill, Birkinshaw, "Ambidexterity and Survival in Corporate Venture Units", Journal of Management 40(7), 1899–1931 (2014): 95 corporate venture units surveyed in 2001, 22 (23 %) closed by 2003; relationships with parent executives and business units raise ambidexterity and support survival — doi.org, open access
- Kinni, interview with Vijay Govindarajan, strategy+business, 14 January 2011 (on Govindarajan, Trimble, The Other Side of Innovation, 2010): "There has to be a link"; make the link, expect the conflict, manage it — strategy-business.com
- Global Corporate Venturing, poll on corporate investors' board seats (2022, updated 2023, n = 134): 90 % hold some board position — globalventuring.com
- McKinsey, "CEOs' choice for growth: Building new businesses" (November 2023), survey of 1,010 executives; self-reported, consultancy survey: C-suite sponsor and regular meetings with core leaders among the practices of the most successful builders; five or more of twelve mechanisms associated with a 2.4 times higher success rate — mckinsey.com
- DellaVigna, Malmendier, "Paying Not to Go to the Gym", American Economic Review 96(3), 694–719 (2006): 7,752 members of three US clubs; members on monthly contracts above 70 dollars attend 4.3 times a month on average, paying more than 17 dollars per visit against 10 dollars with a ten-visit pass — doi.org
- Ashraf, Berry, Shapiro, "Can Higher Prices Stimulate Product Use? Evidence from a Field Experiment in Zambia", American Economic Review 100(5), 2383–2413 (2010): higher prices screen out those who use the product less; no consistent evidence of sunk-cost effects — doi.org
- Network World, "Cisco unapologetic on spin-in model" (2013); Fortune, 7 June 2016: internal resentment over the Insieme spin-in, departure of the founders — networkworld.com, fortune.com
