Strong returns alone won't stop a venture unit from being shut down. The units that make it through put as much work into the people around them, as in their deals.
Most innovation leaders know how the story goes: a venture unit launches with big promises and a CEO announcement on an earnings call. A few years later, the updates slow, the team moves into M&A or strategy, and the unit’s name drops off the organizational chart. Often, the unit had delivered credible investment results.
Xerox has lived that story - twice. In 1989, the company launched Xerox Technology Ventures, a $30 million fund that backed startups such as Documentum and grew past $200 million in seven years. Xerox closed the fund in 1996, as internal conflict grew over who owned the gains and who deserved the credit. In 2021, the company tried again with Xerox Ventures, and a few years later it closed that unit down and moved its investments into an independent fund.
Xerox’s fund performed well, but internal conflict still brought it down. Seven years of research suggests that this occurs in many corporations. In a recent Harvard Business Review article, Ezra Carlson and Mehdi Safavi of Cranfield School of Management, with Nicolas Sauvage, president of TDK Ventures, share findings from their work with more than 100 participants from corporate venture capital (CVC) units, which invest in outside startups. Although Xerox sat outside their study, the researchers use it as their lead example of how stubborn internal tensions can be. Like Xerox, the units in their study faced their biggest threats from inside the company. Most importantly, the ones that survived built routines to manage those internal tensions.
Each routine below gives innovation leaders a check to run against their own unit, before the next budget review.
1. Run the first pilots with business units open to experimentation
In its first year, a venture unit has to choose which business leaders to work with first. Many units start with their biggest skeptics. In the researchers’ experience, that choice often backfires. A young unit has scarce attention and needs to show value quickly.
The units that lasted started with the leaders most open to experimentation. Each unit ran early pilots and co-investments with those leaders, produced a few visible wins and shared the lessons across the company.
Those first partners became advocates. They could explain, from their own experience, what the venture unit added. Their support made other business units easier to win over later.
2. Write one charter with the CEO and CFO
Wider support brings a new problem into the investment committee: each stakeholder expects something different from the unit. One wants a financial fund, another wants a sourcing arm for its business unit, and a third sees a branding signal. As a result, each committee meeting turns into a debate about why the unit exists.
The units that perform well settle that question in a short charter written in plain language. The CEO, the CFO and key business-unit leaders write the charter together with the venture team. These charters usually balance strategic learning and option-building with risk-adjusted financial returns.
Leaders then point to the charter to explain each decision. When disagreements about purpose come up, the committee handles them openly, using the charter as the reference. The same argument stops repeating with every deal.
Checklist: the one-page charter.
For venture unit heads and their sponsors, to complete before the next investment committee or annual review. It gives every later decision a shared reference point. Answer three questions with your CEO, CFO and one or two business-unit leaders:
Why does the unit exist?
What do good results look like?
What falls outside the unit’s job?
3. Name the bridges and map the path to a pilot
A charter agreed at the top needs people in the business to put it to work. On the organizational chart, a venture unit may sit in strategy, finance or corporate development. Its results depend on the people who link it to the rest of the company.
Alphabet’s venture arm, GV, built those links from the start. GV, which sat outside the study, runs its own fund, investment committee and pay system. To stay close to Google, GV shares strategic themes with the parent, meets regularly with product and business leaders, and follows clear rules for when portfolio companies connect with Google. Since its launch in 2009, GV has backed startups including Uber, Slack and Stripe, and it has stayed active through several market cycles.
The units in the study build the same links, at two levels:
Executive level: a small group of senior and business-unit leaders hold standing seats on investment or advisory committees.
Working level: business-unit managers join the unit on secondment, technical experts co-lead pilots, and a few people act as go-betweens for startups and internal teams.
The go-betweens explain each side to the other, so they catch small misunderstandings early, before those misunderstandings grow into crises.
However, even with the right people in place, many collaborations stall over messy processes, slow approvals and confusion over who owns a pilot. The strongest units map each stage from first contact to live collaboration. They add standard templates, clear handoffs and defined roles, then keep improving those steps over time. As a result, the team treats each delay as a lesson for the next pilot.
Exercise: the bridge list.
For venture unit heads at the start of a planning cycle. It turns informal goodwill into named roles and shows where the path to a pilot gets stuck.
List 5–10 people who could connect the unit to the rest of the organization.
Give a few of them an explicit role, such as nonvoting committee member or pilot sponsor.
Set a regular touchpoint with each one.
Map the journey from first startup contact to live pilot, and pick one bottleneck to fix next quarter.
4. Agree the rules on risk and time with legal and finance
Smoother handoffs bring more pilots to legal, compliance and operations leaders, who often have good reasons to worry about experiments. When the rules for testing are vague, a veto is the safest answer for those risk owners.
The units in the study worked with the risk owners to define the conditions for testing, which were:
sandboxes
pilot frameworks with guardrails
joint review forums that involve compliance early
With the rules written down, risk owners find it easier to approve experiments.
Approved pilots then reach a second gatekeeper: finance. Venture units often get measured with the quarterly targets and short-term ROI built for the core business, while the value of their work can take years to emerge. The units that coped best agreed with finance and strategy on three separate horizons: learning, options and financial returns. Alongside cash returns, those units tracked validated insights, capabilities tested, and strategic doors opened or closed. As a result, leaders could weigh short-term pressure against long-term progress.
5. Tell the unit’s story every quarter
Long-term measures help leaders judge a venture unit, and perceptions decide who works with it. Inside the company, the unit can look like a pet project or a black box. Outside, founders can see it as slow or unpredictable. Those perceptions shape whether business units engage and whether founders take the first call.
The successful units manage their reputation as part of the job. They share short, honest stories about wins, failures and lessons, often through people who have rotated between the unit and the business. The researchers suggest a simple rhythm: a short update each quarter, one internal case example and one founder testimonial, with honest messages the unit can repeat.
Over time, the unit becomes known as predictable and open. That reputation earns the unit trust when projects take longer than planned.
