Contents08 sections
The scouting team of an industrial group finds an interesting startup. A good team, a working product, the right market. The file goes to M&A. M&A has little time to decide whether to kill it. Most of the time they kill it.
The deal flow gets thrown over the fence. Whether the right deal survives comes down to chance.
Scouting finds what exists. The question is what you do when the thing you need doesn't exist.
Outside-in: three hopes in one deal
Scouting, corporate venture capital and venture clienting all work the same way: outside-in. Someone else founded a company, and you check whether it fits. That's a sensible approach when the answer to your problem is already out there.
But every outside-in deal rests on three hopes. That the startup fits your strategy. That it still needs you once it's grown. And that it doesn't pivot away from your market at the next financing round. None of these is in your hands.
That's why corporates tend to come in late, once the business model is stable and the hopes have turned into facts. The data shows it. According to the PitchBook-NVCA Venture Monitor, corporate investors took part in 21 percent of US venture deals in 2025, but in 63 percent of deal value. Few deals, big tickets. Late money buys certainty. And certainty has its price.
The CVC that arrives when the problem is already there
Song Ma looked at why companies set up venture units in the first place. His finding in the Review of Financial Studies (2020): companies start a CVC unit after their own innovation has weakened. And they shut it down again once it recovers. The median unit lives around four years.
That's not an accusation. It's a pattern. The venture arm is often the answer to a gap that has already opened. And it searches the market for someone who happens to have closed it.
Sometimes that works. Sometimes the timing is right and the startup exists. Sometimes a challenge put into the right channel finds someone who already solves it.
And sometimes it doesn't.
Buying isn't free either
Even when the right startup exists, the deal is only the beginning. Daniel Kim analysed around 4,000 acquisitions of US high-tech startups using US Census data. 33 percent of acquired employees leave within the first year. For comparable regular hires it's 12 percent.
You pay for a team, and a third of it is gone after a year. You pay a price set by a market that ran up the valuation for years. And you integrate a company that was never built for you.
Fair is fair: scouting works
Now the counter-argument, and it deserves to be taken seriously. Chemmanur, Loutskina and Tian showed in 2014 that startups backed by corporate investors produce significantly more patents, before and after their IPO. Dushnitsky and Lenox found back in 2006 that CVC creates value for the parent company too, especially when it's used as a window on technology rather than as a return engine.
Neither of these contradicts the thesis. They sharpen it. A window shows you what exists. It doesn't fill a gap where nothing exists. And the value is created for the startup, not necessarily for the strategy of the company that invested.
Reverse scouting: start from the gap
That's why I think it needs a second direction. Not instead of scouting, but alongside it. Reverse scouting starts at the other end: with your strategy and your problem, not with a pitch deck.
The question then isn't "who out there fits us?" It's "which company would we buy tomorrow as an M&A target, if it existed?" And then you build exactly that company. Outside your own walls, with your experts, staged gate by gate, with a pre-agreed right to buy it once it works.
It sounds unusual. It isn't. Cisco did exactly this, several times. The SEC filings show the mechanics plainly. Andiamo was funded outside Cisco from 2001, Cisco held an option to buy it at a negotiated formula, and in 2004 it took the company over at an estimated value of 750 million dollars, derived mainly from a revenue-based formula. Nuova and Insieme followed the same pattern: built outside, with a put/call agreement and milestone payments based on agreed formulas. All three became established Cisco product lines: MDS, Nexus and ACI.
The model also had its dark side. Inside Cisco there was envy of the spin-in veterans, and the question of why shareholders were paying hundreds of millions for products they had essentially already financed. The lesson for me isn't "hands off". It's: fix the formula on day one, make it transparent, and don't build in a way that leaves half the organisation feeling passed over.
Build, borrow, buy: most choose just one
Laurence Capron and Will Mitchell studied how companies close capability gaps: build, borrow through partnerships, or buy. Their core finding: most companies misjudge the gap and reach for the route they always take. In their telecom study, only a third actively used more than two routes.
That's exactly the trap. Whoever has only a scouting team only scouts. Whoever has only M&A only buys. And where neither has an answer, the topic ends up on the "later" list.
Two halves of one portfolio
In practice the two directions lock together. Every build leaves gaps that scouting can fill: a partner for a missing piece, a startup worth a stake. And every gap the scouts can't fill becomes a build candidate. Outside-in and inside-out aren't rivals. They're two halves of one portfolio.
For the scouting team that's good news, not competition. They know the market better than anyone. They're the door, not the obstacle. Reverse scouting just gives them an answer for the case where their search comes back empty.
What that means concretely:
First: start with the gap, not with the deal flow. Which piece is missing from your strategy that you'd buy tomorrow if it existed?
Second: check the market honestly. If someone already solves it, partner, invest or buy. That's the faster route.
Third: if nobody solves it, build it. Outside your structure, staged, with a formula for the way back agreed before the first euro flows.
Fourth: link both directions. One portfolio, one board, one list. No idea gets lost because it came from the wrong side.
Don't wait for the right startup to walk through your door. Build the M&A target of your future.
Sources
- PitchBook-NVCA Venture Monitor Q4 2025, p. 23: share of US VC deals with CVC participation 21.3 % by count, 63.4 % by value (2025) — nvca.org
- Ma, "The Life Cycle of Corporate Venture Capital", Review of Financial Studies 33(1), 358–394 (2020): CVC units are started after internal innovation deteriorates and shut down when it recovers; median life cycle around four years — doi.org, author summary, Harvard Law School Forum
- Kim, "Predictable Exodus: Startup Acquisitions and Employee Departures", working paper, US Census matched employer-employee data, around 4,000 acquisitions 1990–2011: 33 % first-year departure of acquired workers vs 12 % of comparable hires — MIT Sloan, census.gov
- Chemmanur, Loutskina, Tian, "Corporate Venture Capital, Value Creation, and Innovation", Review of Financial Studies 27(8), 2434–2473 (2014): CVC-backed firms more innovative than IVC-backed firms — Harvard Law School Forum
- Dushnitsky, Lenox, "When does corporate venture capital investment create firm value?", Journal of Business Venturing 21(6), 753–772 (2006): value creation strongest when CVC is used as a window on technology — ideas.repec.org
- Cisco Systems, Form 8-K, Exhibit 99.1 (2006): Andiamo funding, call option at a negotiated formula, estimated value of 750 million dollars primarily derived from the revenue-based formula — sec.gov
- Cisco Systems, Annual Report FY2008, Note 3: Nuova, milestone payments based on agreed-upon formulas — sec.gov
- Cisco Systems, Form 10-Q, quarter ended 26 October 2013: Insieme, put/call option agreement, milestone payments based primarily on revenue — sec.gov
- Vance, The Register, 11 August 2006: internal backlash against the spin-in model — theregister.com
- Rohrbeck, Döhler, Arnold, "Combining spin-out and spin-in activities – the spin-along approach", ISPIM 2007 — mpra.ub.uni-muenchen.de
- Capron, Mitchell, Build, Borrow, or Buy, Harvard Business Review Press (2012); INSEAD Knowledge: companies tend to misjudge the capability gap; only a third used more than two modes — knowledge.insead.edu
