TL;DR
Roughly a quarter of McKinsey's global fees are now tied to outcomes, with time-based billing covering the restWall-Street-Journal-Berichterstattung zu McKinseys ergebnisgebundenen Honoraren, Fundstelle AI Weekly, and Deloitte has shown its own consultants a projection in which the billable hour shrinks to a sliver of the market by 2035Wall-Street-Journal-Bericht zur internen Deloitte-Präsentation über die abrechenbare Stunde, Fundstelle The Decoder. Meanwhile, 43 percent of consultancies in Germany say AI hasn't affected their pricing at allLünendonk-Liste 2026 Managementberatung (68 Häuser), Bericht auf consulting.de vom 02.07.2026. Software already lists outcome-based pricing on the rate cardHubSpot Company News: Preisumstellung für Customer Agent und Prospecting Agent zum 14.04.2026. If you sell or buy professional services, price at least one service module against that benchmark in 2026.
One gap in the data has kept me busy for weeks. The two ends of the consulting market are telling different stories about the same technology, and only one of them can be right.
At the top, the rebuild is documented. The Wall Street Journal reports that about 25 percent of McKinsey's global fees now come from outcome-based arrangements, with the rest still billed on time and effortWall-Street-Journal-Berichterstattung zu McKinseys ergebnisgebundenen Honoraren, Fundstelle AI Weekly.
In June, Deloitte's US public sector leadership showed staff a chart in which hourly consulting work shrinks to a thin sliver of the total market by 2035Wall-Street-Journal-Bericht zur internen Deloitte-Präsentation über die abrechenbare Stunde, Fundstelle The Decoder. One consultant summed up the internal reaction: they heavily implied the model is toastWall-Street-Journal-Bericht zur internen Deloitte-Präsentation über die abrechenbare Stunde, Fundstelle The Decoder.
Across the broader market, the picture is calm. In Lünendonk's survey of 68 management consultancies in Germany, 43 percent report no effect of AI on their pricing whatsoever, and just 19 percent have raised prices, citing better qualityLünendonk-Liste 2026 Managementberatung (68 Häuser), Bericht auf consulting.de vom 02.07.2026. Clients, for now, aren't claiming the AI dividendLünendonk-Liste 2026 Managementberatung (68 Häuser), Bericht auf consulting.de vom 02.07.2026.
The middle holds still while the top rebuilds
That gap has a short half-life. Fee models tend to look stable right up until a rebid, and procurement teams are already signing software contracts where the outcome sits directly on the price list. I traced how fast those billing units are moving in SaaS pricing in the agent era. HubSpot has charged 0.50 dollars per resolved customer conversation since April, down from 1.00 dollar per conversation regardless of outcomeHubSpot Company News: Preisumstellung für Customer Agent und Prospecting Agent zum 14.04.2026. Among AI product companies, 18 percent already bill on resultsICONIQ Growth: State of AI 2026, Bi-Annual Snapshot (rund 300 Befragte). Three out of four software companies changed price or packaging within the last yearGrowth Unhinged: The 2026 State of B2B SaaS and AI Monetization (230 Unternehmen, April/Mai 2026).
Consulting buyers sign both kinds of contracts. It's a matter of time before they ask why one invoice guarantees a result and the other one bills attendance.
| Criterion | Day rate (time and material) | Fixed fee | Outcome-based pricing |
|---|---|---|---|
| Price anchor | Days worked | Defined scope | Measured result |
| AI efficiency gain goes to | The client (fewer days) | The provider (lower cost) | Both, depending on calculation |
| Provider incentive | Keep utilization high | Keep scope tight | Deliver the result |
| Prerequisite | Trust in timesheets | Stable scope | Measurable unit plus baseline |
| Typical failure mode | Punishes speed | Change-request disputes | Disputes over measurement |
The day rate prices the wrong unit
The day rate prices effort, and effort was a workable proxy for value as long as the two moved together. When a market analysis took four weeks, 20 billable days on the invoice needed no defense.
AI breaks that link. If the same analysis takes a third of the time, a provider has three options. Bill fewer days and shrink revenue. Bill the same days and hope nobody checks. Or change the unit and get paid for the result. The 19 percent raising prices on quality groundsLünendonk-Liste 2026 Managementberatung (68 Häuser), Bericht auf consulting.de vom 02.07.2026 are the visible start of option three.
Anyone who has priced proposals in day rates and steered teams by utilization knows how deep this system sits. Staffing, forecasting and partner economics all hang on sold days. A fee model is the operating system of the firm behind it.
What does outcome-based pricing look like in consulting?
Outcome-based pricing means a predefined, measurable unit of result gets paid, and the time it takes stays the provider's problem. It's simpler than value-based pricing, which negotiates a share of total value created, and narrower than a success fee tied to a deal. It is a unit price for a unit of impact.
Most consulting fields have such units; they're just rarely sold that way. A completed supplier audit. A passed certification. A qualified, accepted lead. A pricing change implemented with a measured delta against baseline.
The unit does real contractual work. It defines what "done" means, it names the baseline the result is measured against, and it decides who carries the risk of a miss. That's also why the conversion effort is front-loaded: the hard part is agreeing on a measurement both sides will still accept in a dispute.
A workable baseline fixes the starting value the impact counts from and the window it's measured over. It also names who runs the measurement. A delta without an agreed window invites cherry-picking from both sides.
A quick illustration (example calculation): a documentation project across four sites costs 20 consultant days at 1,800 euros, so 36,000 euros. With AI tooling the provider finishes in 12 days. On a day rate, the invoice drops to 21,600 euros and the entire efficiency gain lands with the client. Priced per outcome at 9,500 euros per audit-ready site, the same project bills 38,000 euros. The client pays for a guaranteed result, the provider keeps the return on its own tooling investment.
The realistic path is hybrid. In software, 37 percent of vendors now run hybrid models, a base fee plus usage or outcome components, up from 25 percent a year earlierGrowth Unhinged: The 2026 State of B2B SaaS and AI Monetization (230 Unternehmen, April/Mai 2026). Translated to consulting, that's a retainer for availability and governance plus an outcome component for the defined unit of impact.
How do you pick the right fee model for a given service?
Three test questions settle the choice, and they work from both sides of the table.
- Can the result be measured cleanly and attributed to the work? "Procurement cycle time from 12 days to six" is a unit, unlike "a more modern organization". Without a baseline, every outcome invoice becomes an argument. If the answer is no, stay on day rate or fixed fee and build measurability first; that alone is a billable work package.
- Does the provider control the variables that drive the result? A certification project can fail on missing client documents, a lead program on a sales team that lets handovers sit. If the outcome depends on the client's data, decisions or deliveries, write a cooperation clause into the contract or go hybrid: base fee plus outcome component.
- Will the cost base survive a miss? Outcome-based pricing shifts delivery risk to the provider, so the production cost per outcome unit has to be a known number. For AI-supported services you can put a figure on it with the AI cost calculator. Selling outcomes without that number is selling blind.
Then convert exactly one service module and leave the rest of the portfolio untouched. Three months with a real pilot teach more about baselines, measurement and dispute points than any internal model ever will.
What changes inside the firm when the unit moves
A pilot exposes the internal side as well. Firms steered by utilization need different controls for outcome-based pricing: the cost per outcome unit and the rate at which a module actually reaches its result. Forecasting moves from sold days to expected units, and sales needs an answer for the case where the result fails to land.
For one service module the rebuild stays manageable; a cost sheet and a risk buffer in the price will do at the start. Sequence beats speed here: measurement first, then the price. The pitch comes last, because the other way around sales ends up selling a promise production hasn't costed yet.
My Take
The 43 percent look like an incentive problem to me: as long as utilization is the internal steering metric, the day rate stays rational for every individual partner, right up until a competitor prices the result and the rebid flips.
That's why I expect the rebuild in Europe to be forced from the buy side, by PE portfolios and procurement teams that already know the price per resolved case from their software stack. Buying habits are slow, granted, possibly slower than I'd like.
Put a date on it and I'd say the first major DACH tender requiring an outcome component in the fee model lands before the end of 2027, out of a PE portfolio.