Value creation without multiple expansion: which levers still carry PE returns

What drives returns when multiples stop expanding?

Operating improvement does. When the exit multiple is flat or below entry, equity value grows only if EBITDA grows, if cash pays down debt or funds growth, or both. McKinsey's global private markets report says multiple expansion and cheap leverage, which together accounted for 59 percent of PE returns between 2010 and 2022, have faded, so operational value creation is now likely the primary source of returns.

That moves the operating partner from supporting the thesis to carrying it. Each lever has to be specific, owned by a named executive at the portfolio company and visible in numbers that survive a buyer's quality-of-earnings review.

How much harder is the return math now?

Considerably harder. Bain's Global Private Equity Report 2026 estimates that a deal which needed 5% annual EBITDA growth a decade ago to reach a 2.5x return over five years now needs about 12%. The same report puts buyout holding periods at exit at around seven years, up from an average of five to six years in 2010-2021, and says sponsors are holding assets longer to buy time to grow EBITDA.

Which operating levers still work?

The levers that work are the ones a CEO can own and a buyer can verify. For each one, ask who owns it, when it reaches the P&L and what evidence will survive diligence.

1. Pricing and packaging. Often the fastest margin lever, especially where prices have not been reviewed since acquisition. Evidence: realized price by customer cohort and churn after increases.
2. Commercial effectiveness. Pipeline discipline, win-rate improvement and account expansion. Evidence: several years of CRM history showing conversion by stage.
3. Procurement and cost. Vendor consolidation and contract renegotiation. Evidence: run-rate savings backed by signed contracts.
4. AI and automation in operations. Support, finance and document-heavy work. Evidence: cost per transaction before and after, not pilot demos. The guide on why portfolio company AI pilots stall explains why many never get this far.
5. Add-ons that are actually integrated. Evidence: combined systems, retained customers and realized synergies.
6. Cash and working capital. Evidence: days sales outstanding, inventory turns and payables terms.
7. New income from existing assets. Licensing intellectual property, spare capacity or operational records. Evidence: signed agreements and cash received.

The playbook on AI value creation in private equity goes deeper on the fourth lever. An AI disruption risk assessment for the portfolio helps separate companies where AI is a defensive priority from those where it is upside.

Four rules for a plan that expects no help from the multiple

• If a lever cannot name an owner at the portfolio company, it is a hope rather than a lever.
• If its effect will not show in reported numbers before the sale process starts, count it as support for the exit story, not as EBITDA.
• If it is one-time, such as a license fee or an asset sale, model it as cash rather than run-rate earnings.
• If it needs the sponsor's own people to do the work after kickoff, confirm the operating team actually has the capacity.

Where does a one-time data license fit?

It fits as a non-dilutive cash event, not as a growth story. A qualifying company licenses a defined set of its own records, such as support tickets, project files, email threads or engineering history, to AI developers for an agreed term. The company keeps ownership, approves scope and price, and receives one all-in payment, typically within about 60 days of invoicing once the buyer selects the data.

Three practical points for the plan:

• Treat it as non-recurring. Expect a buyer's quality-of-earnings work to normalize a one-time license out of adjusted EBITDA. Its value is the cash, not the multiple.
• Ask the auditors early. How a license is structured can affect when revenue is recognized; Deloitte's ASC 606 licensing guidance explains the difference between a right to use and a right to access intellectual property.
• Check the credit agreement. Covenants can limit how proceeds are used or moved out of the company, so the CFO and lenders' counsel should confirm before the cash is earmarked.

This is general information, not legal, tax or financial advice. Confirm accounting and covenant questions with the company's auditors and counsel.

Not every company qualifies. The who qualifies page sets the bar. The company must be US-based, must have had 50+ full-time employees at peak (contractors excluded), needs a multi-year operating history recorded across many systems, must hold the right to license those records and needs an executive willing to sponsor the process. Licenses are typically exclusive for AI training over an agreed period, so the board should weigh that before signing.

What this means for operating partners who refer companies

Operating teams are larger and more accountable than they were. The same McKinsey report finds firms have more than doubled their operating groups since 2021, and 53 percent of 300 surveyed LPs ranked a GP's value-creation strategy among their top five selection metrics. A lever that can be documented company by company strengthens that story, even when it is a small one.

For the operating partner, the work is the introduction. SourceX qualifies the company, works through the data inventory with management, agrees price and terms with the company, takes the opportunity to AI labs and data buyers, and runs contracting and delivery. 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, payable only after the buyer pays and SourceX receives its fee. That reward comes out of SourceX's share, so the portfolio company's proceeds are untouched. Before you start, check your firm's policy on compensation connected to portfolio companies and ask compliance whether the fund documents treat such a payment as something to disclose to LPs or offset; then read how a sponsor should introduce portfolio CEOs to an outside program before raising it.

Limits and open questions

• Scale. A license will not close the growth gap Bain describes. Treat it as a supplement to operating work, never a substitute.
• Timing. Demand varies by data type and moment. Once a company is deal-ready, buyers typically respond within about two weeks, but no outcome is certain.
• Exclusivity. A future acquirer may ask about any AI-training exclusivity still running, so disclose it in the data room.
• Rights. Records that belong to clients, or consumer and health data without a licensing basis, are out of scope.
• Averages. The McKinsey and Bain figures describe the market as a whole. Your own return math depends on entry multiple, leverage and sector.

Next step

Put each company's levers on one page and add a line for income from existing assets. Where a company meets the baseline, register as a partner and introduce it, or use the referral earnings calculator first to see how the published reward formula works.