Why Corporate Incubators Fail
And an idea on what to do about it
My name is David Weekly. I’ve been in Silicon Valley for nearly 25 years now, invested in 60+ startups, created mexican.vc (returning >11x cash-on-cash), founded three companies and two non-profits. I’ve created corporate R&D teams at both Facebook and Google presenting our team’s work to the CEO, and seen a huge amount of innovation happen everywhere from garages to emerging markets to big, publicly traded companies. There are a lot of ways new ideas can be had, explored, vetted, and brought to market. But one model in particular seems consistently less likely to yield exciting outputs: the corporate incubator.
By “corporate incubator” I mean here a new team within a large company that is tasked with coming up with an idea for a wholly new business and bringing it to life, mostly outside of the aegis of an existing product suite.
Failure by companies to create new companies seems at first glance quite puzzling. There are incredibly smart and hardworking people who work at big companies who are coming up with new and useful ideas all the time. Large companies have tremendous resources to put behind an idea — time, money, people, facilities, testing, and go-to-market channels and relationships. They also already have backoffices for taking care of mundane things like HR, contracting, vendor management, hiring, performance management, sourcing, procurement, accounting, etc. In theory, it should be the best place to realize powerful new businesses.
In practice, it’s not.
My first personal experience with a corporate incubator was back in 1997 doing consulting work for Casio US R&D in Silicon Valley under the gentle guidance of Dr. BJ Fogg, a Stanford social scientist famous for his work in “captology” and behavior design. As a college undergrad, it was improbably idyllic; we’d sit on the lawn and muse about the ways in which music players were going to become small; they’d be solid state and robust and hold thousands of hours of music. You’d be able to listen to any song you wanted anywhere you wanted like an invisible orchestra following you around. And we were thinking about what kind of a world that would imply; alarm clocks you could talk do and could talk back. It’s reality in 2020 but it was fantasy in the 90’s. We came up with all sorts of stuff — and none of it got built. Casio had put themselves into a bind where they needed the US team to be arm’s length enough to be thinking completely differently from the Japan team, but that meant that ideas landed back in HQ with no champion. So none of it got built. Eventually Casio realized they weren’t getting good return on their investment in a fancy Silicon Valley campus with very little to show for it so they closed it all down. The problem wasn’t that there weren’t any ideas or that Casio couldn’t have executed on any of them. The people doing the inventing weren’t able to become the people bringing the thing to market.
Two Awkward Outcomes
The problem is that a corporate incubator will generate, roughly, two classes of ideas — business concepts that are close to what the parent company works on and concepts that are not. Both are doomed.
For an idea that is close to what the company already does, it will either compete with the company’s existing line of business, in which case some VP will burst into the CEO’s office and demand it be killed (I saw this happen at Google and torpedo an Area 120 project that was slightly too promising), or it will be rational complimentary part of a product portfolio, in which case the VP for that area will suddenly have a possibly not-wanted or out-of-order project on their hands. This results in a “graduation” out of the incubator to within the product area…but an awkward/forced landing. Such projects live at the mercy of their executive sponsor and the degree to which it’s a fit with the rest of the product portfolio and almost not at all based on the degree to which the market is excited about the product. Consequently internal adoption is made orthogonal from market adoption, which is usually a recipe for failing to come up with new useful things that the market wants except by accident.
For an idea that is outside of the parent company’s core line of business, the parent is now in an awkward position because they have little to offer in the way of expertise, channel, or partnership and any resources they spend on this offshoot are by definition coming at an opportunity cost of continuing to productively invest in the core business. It’s a huge distraction. Only a handful of companies are large enough and with big enough margins to be willing to fund unrelated companies for an extended period of time. And the parent company is pretty strongly disincentivized to properly fork off the endeavor — to make its product lead the CEO, give it a round of funding, and let it go its own way as an independent company. For many companies the point of creating an incubator wasn’t to offer an exit package for their most innovative and best-connected employees, nor was it to part with valuable cash in exchange for equity in a bunch of random early-stage startups. There are more efficient vehicles for the latter, anyhow.
This gets to the real rub, which is that most parent companies that start an incubator don’t have a solid rationale for why they’re doing so. It’s often tied to brand values — wanting to be seen as innovative and cool. To have a halo to let them attract and retain top talent that might otherwise be at their own startup. And, perhaps, a humility that it’s better to obsolete yourself than to have another company do it. But these high-level, unobjectionable talking points that can coax a Board out of an investment are different than actually knowing to do with promising teams on your hands that are going somewhere.
Failure is easy to deal with; “we tried a lot of exciting new stuff, we learned about a bunch of new spaces and it’s informed our approach as a company” has a non-embarrassing ring to it. Success is the awkward outcome here that nobody seems to plan for. Wryly, it’s a good thing that that doesn’t happen all that often.
What are the “emotional design” factors that limit the likely success of a corporate incubator?
When you are starting your own company, the failure case is pretty bad — you are not going to get paid anything, it sucks to go on a job search, and you’ll have to explain that block on your resume a hundred times; so there’s a great deal of incentive to figure out a viable business.
If you are at a cushy corporate incubator job, there’s a good chance you have great pay and great benefits, and it basically doesn’t matter if you succeed or fail. You’ll get paid just fine. Your family will be okay. There’s no meaningful consequence for failure. Maybe you won’t get promoted as quickly or maybe you’ll have to transfer back to the mothership and work on what shade of button will get users to click on ads 0.31% more frequently but then again that’s probably the team you transferred out of and they’ll be happy to have you back. Besides, Team Margarita Tuesdays were pretty rad.
When you’re at a corporate incubator, you’re highly unlikely to be able to get even a fraction of the upside in value you’re creating if the invention is successful. Very few incubators will let you fork off your company with 80% ownership. Most don’t give you any profit sharing or equity model to personally capture value created — the reason why is that they already took care of your downside and you’re a salaried employee whose job it is to come up with new businesses. Your upside is very limited. Maybe you’ll get to run the idea you brought to life, or maybe an upwardly mobile Director will run with it and have you advise. Who’s to say? You don’t have the control or authority over what happens.
As discussed above, the parent corporation may feel threatened by your invention if it’s relevant or feel uninterested in taking the resources required to spin it out if it is not relevant. Xerox PARC invented the laser printer but sat on it because they were worried it would hurt their photocopier business. They came up with the modern GUI but had no willpower to take it to market so they let Apple steal it. Those inventors didn’t see any personal gain, just the satisfaction of seeing their work deployed by someone.
All of this is not to say good innovation doesn’t happen at big companies — it absolutely does, all the time! But this is typically where either the company has an existential challenge that is core to their business model that needs solving (Amazon nailing delivery logistics & warehouse automation, Apple designing their own silicon), or where the company has top-down committed meaningful effort and expense toward exploring a new line of business (e.g. Apple introducing the Watch, Disney launching Disney+)
For those in a corporate incubator, a further complicating factor is that in a corporate incubator, the ultimate arbiter, your “boss” is the team that runs the incubator. Output products will be designed to impress them. Depending on how the incubator is run, that might have very little to do with whether your concept has found product-market fit.
So how can innovation flourish at a big company? I don’t have all the answers but perhaps I can frame some useful questions:
- If the idea is successful, what resources will be available to it, on what terms, and based on what checkpoints? At what point will the team be allowed to incorporate itself and adopt its own independent governance? Will the team be given enough control and equity to be able to pioneer their own fates? After a split, what company resources will they still have access to? How will non-competes/poaching be handled? What equity and other rights or restrictions — such as preventing licensing IP to a competitor — will the parent company retain?
- If an idea is directly competitive with an existing team or plan at the parent company, how will this be handled?
- How will innovation teams be judged in a way most likely to validate market success?
- If an idea is aligned with an existing team or plan, what compensation or benefit should the team enjoy for accelerating that team or plan?
- Can the team hear from potential internal stakeholders with key unsolved problems they have and how they would value a solution? Will teams be able to sell their solutions to internal customers and charge those other areas?
- Can the team utilize the parent company’s partnerships? Sales channels? Sales team? What’s off limits?
- If a team fails to produce a meaningful output in a fixed window of time, what are the consequences?
- Has the parent company built relationships with external venture capitalists and angel investors in a way where promising teams could pitch and receive term sheets to invest in their solutions?
- If a team lands on an opportunity that is cash-flow positive but not venture financeable and the team wants to move forward, will the company support that?
- Are there “forbidden” markets, customers, or products that you do not want teams to work on? Those should be highlighted as early as possible.
Idea: Take The Plunge
One model that I haven’t seen tried is to take the “entrepreneurship” part possible — let employees take a short period to form around teams and ideas or decide to opt-out. After that period comes a big commit and they actually become real startup employees in a startup that has a special relationship with the parent company.
Every team member gets three months as a full-time employee with no responsibility to think. At the end of the three month period, teammates must either rejoin their group at the parent company or “take the plunge” formally leaving the parent company job and entering employment with a newly formed shell corporation that is given a one-time investment from the parent company but has its own CEO and governance. The relationship between the parent company and NewCorp is contractually defined between those two entities and is a part of the seed financing, recognizing that the parent company is paying in “sweat equity” in addition to capital: people, IP, access, potentially office space, etc. To simplify, this deal should be standardized e.g. $300k of capital per employee for 1–5 employees joining in exchange for a 30% equity stake and irrevocable board observer rights.
Founders get a 60% equity stake and there’s a 10% option pool left for future employees. They can decide how to invest the paid-in capital just like a normal startup would.
This clearly forms a boundary line between the parent corp and the child corporations; after budding off the parent corp is contractually bound to honor their relationship with the startup. A VP can’t come along and decide to nuke the team because it’s too close to home.
This model also could play well with the CFO & HR because it can be structured to cap the number of exiting employees and capital investment in a predictable and structured way. There might even be some tax advantages due to the new startup equity being considered QSBS.
I’m not saying this is the right model, but I’d like to see things like this try that give the talented folks working in big companies a chance to strike out on their own with the support of the mothership…and the existing models have all fared so badly it feels time for a change.
I’d love to hear your reactions to this; email me at firstname.lastname@example.org with your experiences and thoughts.
Reflections Two Weeks After Posting (Nov 3)
There were a number of good discussions of this article, in particular on Michael Jackson’s LinkedIn reshare. Overall there were some great points from commenters, some noting successes in corporate incubation (notably Tinder), and some confusing corporate innovation with incubation — it should be obvious that large corporations can be very good at innovation when it comes to driving their core business forward; this article is strictly discussing the incubation of new businesses.
Things that folks have seen work well:
- Strong executive sponsorship: when the CEO is vested in the success of the outcome and will “pull out thumbscrews” to bring VP’s into line if they try to “kill the babies”.
- Arm’s length treatment: the incubees are materially “on their own”.
- Fixed graduation & support structure: offering incubated teams a fixed-length program and pre-defined formula for ongoing support and participation e.g. commitment to maintain a pro rata and/or put up to $XXXk into the startup’s outside-led round. (A variable support structure where the parent corp has to decide how much support to lend and when ends up introducing signalling risk and requires the corporation to have talent on board that can credibly assess startup valuations.)
Response #1: Anonymous
I was emailed this and was given permission to reshare after anonymizing:
I read your “Why Corporate Incubators Fail” and just wanted to say that you’re absolutely right. Unfortunately because I’m running a startup inside $BIG_CORP right now, I can’t comfortably shout that out publicly.
In fact I’ve internally proposed and pushed for the very idea that you propose — a “venture sabbatical”, where funded teams leave the nest with some amount of seed funding and a formal relationship with the incubator. It wasn’t enthusiastically received. I think the core issue is that these incubators are designed for brand halo, employee retention, and internal empire building. (That’s not what they say, but that’s why they’re created and funded). I’ve worked at and with other incubators too, and with the exception of those structured as R&D arms of the core business, these internal incubators are neither measured on, nor created for, financial return. So the venture sabbatical model doesn’t produce a highly-valued outcome for the host organization, and it’s worse for brand halo, worse for employee retention, worse for the incubator founders’ & sponsors political ambitions.
An interesting arbitrage would be for an external seed investor to take advantage of the $BIG_CORP rehire window. An external entity can offer $BIG_CORP a structured “launch a startup” plan with the same 9 month deal you propose; the ex-$BIG_CORP has the comfort of a return-to-$BIG_CORP path if the startup fails to impress at the <1 year mark. It’s not much different from what some incubator programs do today, but you could market it this way.
Response #2: Peter Nixey
A few years back I was part of the Coca Cola Founders programme — http://ww.coca-colafounders.com/ — (lead by David Butler, Ross Kimble and Marius Schwartz). They put together a structure very similar to the one you suggested and actually had great success with it. The most visible success was Wonolo — https://www.wonolo.com/- who ended up being backed by Sequoia.
The idea behind the programme was to take founders from either inside or outside Coke and then to give them resources as an entirely independent entity but to give them access to the problems and the internal markets that exist at Coke.
So Wonolo, for instance, took on the problem of relief staff. One problem that Coke has is that when it’s not on the shelf it doesn’t sell so there’s a lot of talk about how to better provide rush-hour staff to stores who need to restock. This was the seed of the idea that Wonolo ran with.
David, Ross and Marius were an exceptional team. Way more clued up than anyone I’ve ever met in corporate venture (and frankly in venture in general) and extremely humble From what I understood the fund also actually performed very well by any regular standards.
Unfortunately though there was one remaining problem — which was that when Coke started to see profits go down it ended up killing all activities that weren’t central to Coke itself. So even though the fund had done well it got closed down and no new companies received investment. The other problem with that was that the Coke representation on the existing companies ceased to be the original team (since they left). And the people who were left just weren’t interested or motivated to deal with the type or scale of companies that the portfolio represented. The lady who was dealing with me had purchased Costa Coffee three months earlier so I was very, very small fry.
My learnings from that are that I wonder whether for longevity, there is one additional independence criterion to a corporate accelerator. Not only should the companies be independent from the core but the accelerator itself should be independent.
I don’t immediately know how you tie its destiny to that of the host but I think that given that a fund needs about 12 years to show returns and may outlast even the lifetime of the average CEO, it has to be designed to withstand the changing priorities and politics inside the host firm.
My experience with the Coke accelerator was broadly brilliant. It dropped off when it was shut down but I think the premise was great and the team who did it were exceptional.
One final thought. One reason I’ve always thought that trad corporate accelerators fail is the identity problem. Many of the companies that you and I know have been through an extended period of near or actual death before coming out the other side to success. Much of the reason the founders continue is that their identities are so wrapped up in it that they can’t bear the sense of personal failure.
In a corporate incubator where you still give out business cards of the parent company you don’t have that. But if you burned the bridges when people joined, made sure that it’s a separate entity, publicised the founders across the rest of the company and championed them. I think you’d get a much larger incentive for them to push through and give up as late as possible rather than lose face.