# Outcome-Based Pricing for AI-Native Agencies: Rollups and the Measurement Problem

Source: https://trymimetic.com/guides/ai-native-agencies-outcome-based-pricing

[Guides](/guides) / Revenue QA

# AI-native agencies: rollups, outcome-based pricing, and the measurement problem

Agencies are one of the clearest targets for AI rollups. The work is digital, the delivery is repeatable, and AI changes how much one team can produce. The natural endpoint of that pitch is a different way of charging: not hours or retainers, but outcomes. This guide covers who is rolling up agencies, what the case for outcome-based pricing is, and the problem every outcome contract runs into.

## Agencies are being rolled up on shared AI infrastructure

Of the 81 AI rollups we track, a handful focus on marketing and digital agencies. ONAR describes the model plainly: **"We acquire proven agencies and run them on shared AI infrastructure, so they can take on more work without adding people at the same rate,"** and **"Each agency keeps its brand and its experts. What changes is the engine underneath"** ([ONAR](https://onar.com)).

The broader thesis is the same as for every AI rollup. General Catalyst: **"Customers care about outcomes, not products"** ([The Future of Services](https://www.generalcatalyst.com/stories/the-future-of-services), 2025). Sequoia's Julien Bek puts the difference between selling tools and selling work this way: **"A copilot sells the tool. An autopilot sells the work"** ([Services: The New Software](https://sequoiacap.com/article/services-the-new-software), 2026). For the wider market, see [AI rollups in 2026](/research/ai-rollups-2026).

## The case for outcome-based pricing

If the agency sells the work rather than the hours, it can price on the result. The idea was popularized as "service-as-software" by Foundation Capital in April 2024: **"An outcome-oriented approach aligns the cost of the unit of software with the associated business value"** ([Foundation Capital](https://foundationcapital.com/ai-service-as-software/), 2024).

Software companies are already doing it:

- Sierra: **"With outcome-based pricing, Sierra gets paid only when we complete a task for you"** ([Sierra](https://sierra.ai/blog/outcome-based-pricing-for-ai-agents), 2024).
- Intercom's Fin: **"Fin is priced at $0.99 per outcome"** ([Fin pricing](https://fin.ai/pricing)). Intercom says it now carries the risk: **"if we don't perform as well as we think we will, that's ultimately less revenue for us"** ([Stripe case study](https://stripe.com/customers/intercom-pricing)).
- Zendesk announced it was **"First in CX Industry to offer Outcome-Based Pricing for AI Agents"** ([Zendesk](https://www.zendesk.com/newsroom/articles/zendesk-outcome-based-pricing/), 2024).

## What the adoption data says

It is early, and agencies are behind software.

- In Kyle Poyar's 2025 survey of 240+ companies, **"5% of survey participants said their primary pricing model is outcome-based right now. However, 25% said they expect it to be outcome-based by 2028"** ([Growth Unhinged](https://www.growthunhinged.com/p/2025-state-of-b2b-monetization), 2025).
- In agencies, performance pay has gone the other way. The ANA's compensation study found performance incentives used by 41% of marketers in 2022, down from 48% in 2016 and 61% in 2013, while 82% use fees ([ANA](https://ana.net/content/show/id/76981), 2022).
- WPP's CEO Cindy Rose: **"the time and materials model is probably not sustainable in the long term,"** but outcome pay is still years away ([Digiday](https://digiday.com/media-buying/it-will-take-a-few-years-wpp-ceo-cindy-rose-says-outcome-based-pay-is-still-years-away/), 2026). Forrester's Jay Pattissall: **"AI has not as of yet transformed the remuneration model"** ([Digiday](https://digiday.com/marketing/ai-is-scaling-in-advertising-outcome-based-pay-isnt), 2025).

## The measurement problem

Every serious source on outcome pricing lands on the same blocker: someone has to agree what the outcome is and count it correctly.

- The ANA on why agency performance pay stalled: **"marketers still struggle to attribute more specific results to an individual agency partner."**
- AlixPartners: **"Agencies and brands must agree which outcomes count, over what period, and who funds the measurement"** ([AlixPartners](https://www.alixpartners.com/insights/media-entertainment-industry-predictions-report-2027/agency-pricing/)).
- Kyle Poyar's test for whether outcome pricing can work includes attribution: **"Can you convince customers to give your product credit for delivering the outcome?"**
- Bessemer: outcome pricing **"works when you're confident in your AI's performance, when you can absorb cost variance, and when the outcome is unambiguous and measurable"** ([Bessemer](https://www.bvp.com/atlas/the-ai-pricing-and-monetization-playbook), 2026).
- Intercom, on how it judges a successful resolution today: **"we rely on basic rules to track whether a resolution was 'successful.'"**

For a marketing agency, the outcome is usually a lead, a sale or revenue, and those numbers come from the client's website and analytics. That is where outcome contracts break. In our audits of 936 live sites, **one in five had duplicate tracking corrupting its counts**, and **28% had consent settings dropping conversion data** (see [conversion tracking setup](/guides/conversion-tracking-setup)). Ad platforms add their own view: Haus's analysis of 640 Meta experiments found **"platforms are grading their own homework"** ([Haus](https://www.haus.io/blog/is-meta-incremental), 2025).

An agency paid per conversion on a site that double-counts is overpaid. One paid on a site that drops conversions through consent is underpaid. Either way, the contract rewards the tracking, not the work.

## How to make outcome pricing workable

1. **Agree the metric and the source before signing.** Paid's Calvin Field: **"agree on the metric and the measurement methodology before the contract is signed, not after"** ([Paid](https://www.paid.ai/blog/outcome-based-pricing-is-the-unlock-for-full-platform-adoption), 2026). Name which system wins when they disagree: the store, the CRM or the ad platform.
2. **Check the tracking first.** Confirm each conversion is counted once, that consent settings are not dropping it, and that the client's records and analytics agree at a stable ratio.
3. **Measure lift, not totals.** Where spend is large, use holdouts or incrementality tests so the agency is paid for what it caused, not for sales that would have happened anyway.
4. **Use hybrids while it matures.** Foundation Capital expects the shift to be **"messy, with false starts, hybrid tiers, and plenty of margin surprises"** ([Foundation Capital](https://foundationcapital.com/the-4-6t-service-as-software-opportunity-lessons-from-year-one/), 2025). A base fee plus a performance component is where most agencies start.
5. **For rollups: make measurement part of integration.** An agency group that wants to sell outcomes across its brands needs one definition of a conversion and one way of counting it in every client account.

## How Mimetic helps

Mimetic checks a site the way customers use it and whether its tracking adds up: conversions counted twice, consent settings dropping them, tags that stopped firing, and records that disagree with analytics. For agencies and agency groups moving to outcome-based pricing, that is the step before the contract.

## Common questions

**What is outcome-based pricing?**
Charging for a result delivered, such as a resolved ticket, a qualified lead or a sale, instead of for time, seats or a retainer. Sierra and Intercom are well-known software examples.

**Do marketing agencies use outcome-based pricing?**
Some do, usually as a performance component on top of a fee. The ANA found performance incentives used by 41% of marketers in 2022, down from 61% in 2013, mostly because results are hard to attribute to one agency.

**What is an AI-native agency?**
An agency built or rebuilt so that AI does much of the production work, letting the same team serve more clients. Some are being acquired by AI rollups that run several agencies on shared infrastructure.

**What is service-as-software?**
The idea that AI lets companies sell the work itself rather than software to help people do it, and charge for outcomes. Foundation Capital popularized the term in 2024.

**What is the biggest risk in outcome-based pricing?**
Measurement. If the outcome is counted wrong, through duplicate tracking, dropped conversions or platform attribution, the price is wrong too. Agree the metric, the source and the method before signing.

Written by Coleman Hindes, founder, Mimetic.
Published 2026-11-22.

### Check whether a client's conversions are counted once, before you price on them.

[Check a site](/)
