Gain-Share Without Giving Away the Firm

Gain-share pricing pays a provider a share of the value its work creates rather than a fee for the effort it expends. It is the most commercially attractive of the new structures and the one most capable of damaging a mid-sized firm, because the upside is capped by the deal while the downside is capped only by the balance sheet.

Why the industry keeps stalling here

Reporting on Indian IT through 2026 has been consistent about what limits adoption of outcome and gain-share models, and it is not client appetite. Relatively few providers possess the operational depth and financial capacity to assume the risk these contracts transfer. Deals also tend to be smaller and shorter than the contracts they replace, which shortens the period over which a provider can recover an investment.

Infosys chief executive Salil Parekh has described client interest in outcome-based work as real but not yet a large part of activity. The gap between stated direction and booked revenue across the industry is mostly a risk-carrying gap, and the numbers show its width. Cognizant reports roughly 45% of BPO contracts on outcome models; Coforge puts outcome-based work at 6–7% of total revenue. Both firms want the same destination. One has built the capacity to carry the risk and the other is building it.

That is worth stating plainly because it reframes the decision. Declining a gain-share deal is not a failure of ambition. Signing one the firm cannot absorb is a failure of arithmetic.

The four mechanisms that make risk survivable

Floors. A minimum payment regardless of measured outcome, typically covering direct delivery cost. A floor converts a potentially catastrophic contract into a low-margin one. Most first-time gain-share deals should carry a floor at or near cost-to-serve, and a provider unwilling to discuss floors is usually a provider who has not modelled the downside.

Caps. A ceiling on provider upside. Clients ask for these and providers resist, but caps are the price of floors and generally worth paying. A capped, floored contract has a known range. An uncapped, unfloored one has a known range too, and the bottom of it is unpleasant.

Ratchet protection. The subtlest of the four. In a multi-year gain-share, this year's achieved outcome becomes next year's baseline, so the provider must keep improving against its own improvements to earn the same money. Without an explicit reset mechanism, a successful contract becomes progressively harder to deliver profitably until it turns loss-making.

Ratchet protection usually takes one of three forms: a baseline reset at defined intervals, a declining share percentage that acknowledges diminishing returns, or a fixed floor that rises with demonstrated performance.

Dependency clauses. Named conditions outside the provider's control that suspend or adjust the measurement. Client staffing changes, scope additions, data quality below an agreed threshold, adoption rates below a stated level. These are not escape hatches; they are the difference between a contract the provider can deliver and one where the provider is exposed to the client's own decisions.

The downside test

Before signing, the question is not what the contract earns if it works. It is what happens if it does not.

A workable discipline is to model three cases and one decision:

The concurrent case is the one firms skip and the one that matters. A single contract in difficulty is a management problem. Three at once, in a firm of a few hundred people, is a solvency problem.

HCLTech's cloud management agreement with E.ON is instructive at the extreme end. Signed in June 2025, it carries no compensation in the first year, with payment from year two tied to efficiency gains and business outcomes. HCLTech can fund twelve months of delivery from its balance sheet without strain. A provider one-hundredth the size structurally cannot, however good the engagement looks.

Deals worth declining

Some gain-share proposals should be refused politely and early.

• No agreed baseline. If the starting position cannot be measured and jointly signed, the contract has no arithmetic and will end in dispute.
• Outcomes driven by client behaviour. Adoption rates, internal process compliance and business volumes are not provider-controllable. Sharing risk on them is gambling on someone else's management.
• Unbounded exposure. Any structure where the provider's downside has no defined floor should be declined regardless of upside.
• Short term, long payback. A two-year contract requiring eighteen months of investment leaves almost no earning period, and the client holds the option to reprice at the moment it becomes profitable.
• First deal, largest account. Learning should not happen on the account the firm cannot afford to lose.

The upside is real

None of this argues against gain-share. The evidence on the other side is substantial. ExlService reports outcome-based arrangements at roughly 36% of revenue, targeting the mid-40s within three to five years. Genpact, at around 24%, attributes roughly 300 basis points of gross-margin expansion to agentic solutions. Firms that have built the capability are earning more per unit of delivery, not less.

The pattern among them is consistent. They started on bounded contracts, built measurement discipline, learned their real cost of achieving a result, and scaled from there. None of them began by putting a flagship account on an uncapped share of business outcomes.

Risk-sharing is a capability, and capabilities are built in a sequence. The firms that will be taking meaningful commercial risk in 2028 are the ones taking small, well-structured, slightly uncomfortable risk this year.