How portfolio companies diversify revenue, and where data licensing fits

What revenue diversification means for a portfolio company

Revenue diversification is reducing how much of a company's revenue depends on one customer, one product, one channel or one end market. In a lower-middle-market portfolio company, the usual trigger is a diligence finding: a top customer carrying a large share of revenue, one product line carrying the margin, or a sales engine that runs on the founder's relationships.

The goal is not variety for its own sake. A diversified revenue base is easier to finance, easier to sell and less exposed to a single contract renewal. The pressure to find growth has also risen. Bain's Global Private Equity Report 2026 estimates that a deal which needed about 5% EBITDA growth a decade ago now needs about 12% to reach a 2.5x return over five years, so a single growth engine rarely carries the plan on its own.

The four classic routes, and one non-customer route

Most diversification plans pull one of four levers. A fifth, licensing historical records, works differently: it brings a one-time payment from a buyer who is not a customer, without changing what the company sells.

The first four change the business and, if they work, recur. The fifth leaves the business as it is and does not recur by default. A good plan is explicit about that difference.

How to pick a route: the 3D test

Score each candidate initiative on three questions before it goes into the value creation plan.

• Dependence: does it reduce reliance on the current top customer, product or channel, or only add revenue alongside them?
• Distraction: how much management time, hiring and capital does it consume over the next four quarters?
• Durability: will the revenue recur, and will an exit buyer's quality of earnings work count it?

Customer and channel moves usually score well on dependence and durability but poorly on distraction. Data licensing is the reverse: it asks little of management once the inventory is done, because SourceX runs buyer review, contracting and delivery, but it scores low on durability because the payment is one-time. The AI use case prioritization framework applies the same kind of scoring to AI initiatives more broadly.

Where data licensing fits, and where it does not

Data licensing is a cash event that rides on records the company already has. It suits companies that have operated for years across many systems and hold histories of real work: support tickets with resolutions, CRM deal histories with outcomes, engineering reviews, project files, approvals and exceptions.

Buyers want them because AI developers are moving from chat-style models toward agents that complete work, such as triaging a ticket, preparing a quote or reconciling an invoice. Teaching and testing an agent on that work requires real multi-step examples with decisions and outcomes, and those sit inside companies rather than on the public web.

What it is good for:

• Adding cash for debt paydown, an add-on, a technology project or a distribution, without new debt or equity.
• Showing an exit buyer that the company's records are organized and rights-reviewed.
• Testing the depth of the company's records, which also informs its own AI plans.

What it is not:

• It does not fix customer concentration in recurring revenue.
• It is not run-rate revenue and should not be capitalized in an exit multiple.
• It is not a sale. The company keeps ownership, the license is typically exclusive for AI training for an agreed term, and nothing is binding until the company agrees price and terms and signs.

The commercial shape is simple. The company receives one all-in price, with SourceX's fee included and no separate charges, as a one-time payment typically within about 60 days of invoicing once the buyer selects the data.

Which portfolio companies are candidates

Start with the who qualifies baseline: US companies with 50+ full-time employees at peak (contractors excluded), several years of documented operations, rights to license the data, and an authorized sponsor such as the owner, CEO or CFO. Then match the diversification problem to the records the company holds.

The full list of exclusions is in which portfolio companies are not a fit. For a portfolio-wide pass, the portfolio company screening workbook keeps each company's answers in one place.

What it means for an operating partner

An operating partner who spots a candidate makes the introduction and stays out of the data. The sequence after that sits with the company and SourceX:

1. SourceX qualifies size, history, data breadth and rights with the company's sponsor.
2. The company completes a data inventory of its systems and years of history.
3. Price and terms are agreed before anything goes to buyers.
4. AI labs and data buyers review the opportunity; once a company is deal-ready, buyers typically respond within about two weeks.
5. The deal closes, data is delivered under agreed redaction rules, and the company is paid.

Partners earn 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, capped at $100,000 per referred company. The reward is paid only after the buyer pays and SourceX receives its fee, no reward is guaranteed, and it is a share of SourceX's fee, so it never reduces what the portfolio company receives. Check your firm's policies on fees connected to portfolio companies before registering. The operating partner hub and the AI value creation playbook cover the role in more depth.

Limits and open questions

• One-time, not recurring. Plan for a single payment. Nothing in the process creates a contracted follow-on deal, so a plan should not assume a second license.
• Rights decide everything. A company whose records belong to its clients, or are mostly consumer or patient data, may not be able to license at all.
• Demand depends on buyers. Buyers select data against their own training and evaluation needs, so a qualified company may still not be selected.
• Operations now carry returns. McKinsey's Global Private Markets Report 2026 says multiple expansion and cheap leverage, which accounted for 59 percent of PE returns between 2010 and 2022, have faded, making operational value creation the likely primary source of returns. Diversification and licensing both belong on that operating agenda, but neither replaces core growth.

Next step

List the portfolio companies with the heaviest customer, product or channel concentration, then check which of them also hold years of records. The network opportunity finder helps map candidates across your network. To introduce a company that passes, register as a partner; owners can also apply directly at sourcex.si/apply.