Outcome-based pricing means customers pay only when your product achieves an agreed result, such as a resolved support ticket, a qualified lead or a successful hire. Fin charges $0.99 per resolution and Zendesk $1.50 to $2.00 per automated resolution. It aligns price with value, but only works when the outcome is measurable, attributable and defined in advance, usually alongside a base fee or minimum commitment.
Key takeaways
- Outcome pricing charges per result achieved, not per seat or per unit used. It's the far end of usage-based pricing.
- It works best in customer support, where a resolution is countable: Fin $0.99, HubSpot $0.50, Salesforce $2, Zendesk $1.50 to $2.00.
- Use the four tests: is the outcome measurable, attributable, already valued by the customer, and achieved soon after the work?
- Your revenue now depends on your success rate, and you pay for every failure. Model the margin before you launch.
- The safest structure pairs a per-outcome price with a base fee or minimum commitment, plus a written outcome definition.
We'd put more than $200,000 into my first business. It was a "results-guaranteed" marketing agency. Clients paid us when our campaigns delivered results. It was a brilliant way to win work. The company snowballed and started attracting international clients because of that new business model.
So why were we burning cash?
Because every result was a one-off. We'd win a few projects, max out the team, the projects would end, and I'd sit there wondering how to pay everyone. Feast or famine. Our margins were ridiculously low. Eventually one of our best clients bought some of our work and took on my team, and I closed that chapter. (I kept my dignity, if not much money.)
I tell the full story in Chapter One of The Subscription Playbook. I'm telling it again here because outcome-based pricing is back, it's the hottest idea in AI, and the same lesson applies.
Paying only for results is a fantastic way to win customers.
On its own, it's a dangerous way to run a business.
Let me show you how companies like Fin, Zendesk and Sierra are making it work in 2026, the four tests your outcome must pass, and how to structure it so you get the upside without my old bank statements.
What Is Outcome-Based Pricing?
Outcome-based pricing means the customer pays only when your product achieves a result they care about.
Not for access. Not for a seat. Not for how much they use it. For what it achieves.
The best-known example is Fin (formerly Intercom), the AI customer service agent. It charges $0.99 per outcome. An outcome counts when the customer confirms the answer solved their problem, doesn't ask for more help after Fin's last reply, or Fin completes a defined procedure. If the customer asks for a human, Fin charges nothing.
Here's how it compares with its close cousins:
| Model | What the customer pays for | Example |
|---|---|---|
| Seat-based | Access, per person | Microsoft 365 Copilot, $30 per user per month |
| Usage-based | Units consumed | Twilio, from $0.0083 per US SMS |
| Value-based | A price set from the value delivered, charged in any form | Harvey's legal AI, reportedly $1,200+ per seat per month, anchored to lawyer time |
| Outcome-based | Results achieved | Fin, $0.99 per resolution |
In Chapter Four of the book, I called usage-based pricing "the natural evolution of the subscription model". Outcome-based pricing is the far end of that evolution. Usage pricing charges for effort. Outcome pricing charges for success. (If you want the bigger picture first, start with usage-based pricing and value-based pricing.)
Why Is Everyone Suddenly Talking About Outcome Pricing?
AI agents.
When software helps a person do a job, charging per person makes sense. When software does the job, charging per person stops making sense, because every job the software does is a job a person no longer needs a seat for.
Fin's founder Eoghan McCabe put that logic at the heart of the company's bet. If Fin works, customers need fewer human agents and buy fewer seats. Charging per resolution means Fin gets paid for the work, not the headcount. It also changed incentives inside the company: every improvement in the resolution rate earns more revenue, so the team fights for every basis point.
The data shows the shift. ICONIQ's 2026 survey of about 305 AI builders found the share of companies using outcome-based pricing rose from 18% to 23%.
Here's what it looks like across customer support today:
| Vendor | Outcome price | What counts |
|---|---|---|
| Fin | $0.99 per outcome ($1.99 voice, $9.99 sales qualification) | Confirmed resolution, no further help requested, or completed procedure |
| Zendesk | $1.50 committed / $2.00 pay-as-you-go per automated resolution | Resolved without human escalation; 5 to 10 included per agent seat |
| HubSpot | $0.50 per resolved conversation | Customer Agent resolution (50 credits at $0.01) |
| Salesforce Help Agent | $2 per resolution | "No charge for escalations" |
| Sierra | ~$1.50 per resolution (third-party estimate) | Custom success definitions in each contract |
Prices checked 27 October 2026; they change often. See the AI Pricing Index. For more industries, see my list of outcome-based pricing examples.
How Does Outcome-Based Pricing Work?
Every outcome-based price has five moving parts. Miss one and you'll have a dispute on your hands.
- Define the outcome in plain words, including what doesn't count.
- Measure it, ideally inside your own product so both sides see the same data.
- Attribute it. Was it your product, or something the customer did?
- Bill it, usually monthly in arrears, often against a prepaid commitment.
- Handle disputes, with a reopen or refund rule and an audit trail.
Zendesk learned the hard way how much step 5 matters. In May 2026 it added "dual verification of billed resolutions", because customers questioned what counted. Build that audit trail from day one, not after the first angry email.
The Four Tests: When Does Outcome-Based Pricing Work?
Before you change your pricing, run your outcome through these four tests. If it fails any of them, choose a different meter.
Test #1: Is it measurable?
Can you count it automatically, without a meeting? A resolved ticket, a booked appointment, a signed contract, a successful hire. Yes. "Improved customer happiness" or "better brand awareness". No.
Test #2: Is it attributable?
Can you show your product caused it? Fin can, because the conversation happened inside Fin. A marketing tool claiming credit for revenue has a harder job, because the customer's sales team, prices and timing all played a part.
(This was precisely the trouble with my results-guaranteed agency. When a campaign worked, the client's product was great. When it didn't, our marketing was rubbish.)
Test #3: Does the customer already put a dollar value on it?
The best outcomes are ones the customer already tracks in their own reports. Support managers know what a human-handled ticket costs them. Recruiters know what a placement fee looks like. Sierra's pitch rests on outcomes like a saved cancellation, which the customer can tie straight to revenue.
If the customer has to be taught what an outcome is worth, you'll spend your sales cycle on education.
Test #4: Does it happen soon after the work, within your control?
A resolution happens in minutes. A hire happens in weeks. "Customer lifetime value went up" happens in years. The longer the gap and the more people involved, the more you're financing your customer's business while you wait to be paid.
Value Metric FinderScore candidate value metrics and pick the one that grows with your customers.Open the free tool →Worked Example: Why Vendors Moved From Seats to Outcomes
Let's see why outcome pricing is so attractive to the vendor, and what it means for the customer.
A 10-agent support team as the AI gets better
Assumptions (illustrative): a company handles 5,000 support tickets a month with 10 human agents. You sell help desk software at $115 per agent seat per month (Zendesk Suite Professional's annual price), or AI resolutions at $1.50 each. Your model and hosting cost is $0.15 per AI conversation.
Seat model, year one: 10 seats × $115 = $1,150 a month. Seat model after the AI resolves 60% of tickets and the team shrinks to 6 agents: 6 × $115 = $690 a month. Your product got better and your revenue fell 40%.
Outcome model at a 60% resolution rate: 3,000 resolutions × $1.50 = $4,500 a month. Cost: 5,000 conversations × $0.15 = $750. Gross margin about 83%.
Outcome model at a 40% resolution rate: 2,000 × $1.50 = $3,000. Cost still $750. Margin 75%.
That's the vendor's case in one box. Under seats, success shrinks your revenue. Under outcomes, success grows it.
Now look at it from the customer's side.
The bill went from $1,150 to $4,500. That may still be a bargain if six human agents cost far more. But if the resolution rate climbs from 40% to 70% next quarter, the bill jumps 75%, and nobody budgeted for it. The better your product works, the bigger the surprise. (Somewhere, a CFO is sighing.)
That's why so many buyers, given the choice, pick the predictable meter. Decagon offers both, and says the vast majority of its customers choose per-conversation pricing. Salesforce now sells fixed-price Agentic Enterprise License Agreements alongside its per-resolution option, and they grew from 16 to more than 120 in one quarter.
Model your own numbers in the AI Agent Pricing Calculator.
Compare per-seat, per-conversation and per-outcome revenue
Plug in your volumes, success rate and cost per task. The free AI Agent Pricing Calculator shows which model earns more, and what margin you keep.
What Are the Risks of Outcome-Based Pricing?
I'm a fan of outcome pricing done well. But be honest with yourself about the downsides.
Risk #1: Feast or famine, again
Your revenue now depends on your customer's volume and your success rate, two things that can both drop in the same bad month. In Chapter Four I described a telecom customer whose spending dropped off a cliff the month after a busy election. Outcome pricing adds a third variable on top of usage. Plan your cash flow accordingly.
Risk #2: You pay for every failure
Every attempt costs you money. Only successes pay you. If your success rate is 40%, 60% of your costs earn nothing. Watch your cost per attempt as closely as your price per outcome.
Risk #3: Arguments about what counted
The definition is where disputes happen. Was a conversation "resolved" if the customer just stopped replying? Was a lead "qualified" if they never answered the phone? Fin handles this by charging at most once per conversation and refunding reopened cases. Copy that generosity.
Risk #4: Competitors undercut the headline
Outcome prices are simple to compare, so they get compared. HubSpot moved its Customer Agent from $1 per conversation to $0.50 per resolved conversation in April 2026, landing well under Fin's $0.99 and Zendesk's $1.50. If your margin only works at the top of the market, you're exposed.
Risk #5: Slower revenue recognition
Outcome fees are variable, so accounting rules can delay when you recognize them. That matters for your reported revenue and, if you're raising money or selling, your valuation. I cover the basics in revenue recognition for subscriptions.
How to Write an Outcome Definition: A Clause Checklist
This is the part most guides skip, and it's where deals succeed or fail. Before you sign an outcome-based contract, make sure it answers every question below.
Steal this: the outcome definition checklist
- What counts? The exact event, in one sentence, with an example.
- What doesn't count? Escalations to a human, duplicates, test traffic, spam, outcomes the customer caused.
- How often? For example, at most one billable outcome per conversation, per lead or per role.
- Time window. How long after the work can an outcome be claimed, and how long before it's final?
- Reversals. What happens if a ticket reopens, a hire leaves within 90 days, or a booking no-shows?
- Measurement. Whose system is the source of truth, and can both sides see the log?
- Disputes. How many days the customer has to query a billed outcome, and who decides.
- Minimum commitment. The floor the customer pays regardless of outcomes.
- Cap. The maximum monthly bill without written approval.
- Price review. When and how the price per outcome can change.
The Safer Structure: Outcome Pricing Plus a Commitment
So how do you get the upside of outcome pricing without the feast-or-famine downside?
The same way I eventually fixed my own business: recurring revenue underneath.
In Chapter Two of the book, Model #8 is the minimum commitment model. The customer commits to a minimum level of spend in exchange for a better rate. In Chapter Four, I suggested using a minimum commitment to claw back what you spent acquiring the customer, while still giving them the benefits of usage pricing.
That's exactly how the outcome leaders operate:
- Fin sells standalone with a minimum monthly commitment (for example, 50 outcomes).
- Sierra reportedly starts contracts at around $150,000 a year, with custom success definitions and minimum commitments written in.
- Zendesk includes 5 to 10 automated resolutions per agent seat, so the seat fee carries the base.
- Decagon reportedly charges a platform fee of around $50,000 a year on top of conversations or resolutions.
The pattern is clear. A floor, then outcomes on top.
The commitment pays for your onboarding and gives you predictable revenue. The outcome price aligns you with the customer's success and gives you room to grow inside the account. This is a form of hybrid pricing, and for most businesses it's the right answer.
My take
If I could go back and redesign my results-guaranteed agency, I wouldn't drop the guarantee. Clients loved it, and it won us work against much bigger competitors. I'd add a modest monthly retainer that covered our team's core costs, and make the results fee the bonus on top. Same promise to the client. Very different conversation at my kitchen table.
Can Small Businesses Use Outcome-Based Pricing?
Yes, and plenty already do. They just don't call it that.
A recruiter's placement fee is outcome pricing. So is a real estate agent's commission, a "no fix, no fee" repair and a marketing consultant's percentage of sales.
In Feed a Starving Crowd, I described a "lower fee plus percentage of sales" model for consultants, and gave a strict checklist for when to use it. The product should already be selling, the client should have some profile and an audience, their real problem should be execution, and you should have authority to make decisions. That's Test #2 (attribution) and Test #4 (control) in plain English.
If you run a service business, here's a simple way to start:
- Keep your normal monthly fee, perhaps slightly lower.
- Add a bonus per outcome the client already tracks (per booked job, per qualified lead, per lodged claim).
- Define it with the checklist above.
- Review after three months of real data.
If you're an accountant, lawyer or consultant thinking about this, the guide to value-based pricing for services is a good next read.
Australian businesses should also think about cash flow. Outcome fees arrive later and less evenly than a retainer, and depending on how you account for GST, it may be due when you invoice rather than when you're paid. That's worth a conversation with your accountant before you change your terms.
Free Subscription Profit Review
Australian business owners can book a free 30-minute review with a BPAG Accountants adviser. An honest conversation about:
- whether a subscription model could work for your business
- how recurring revenue could cover your fixed costs each month
- how it could lift what your business is worth
No cost, no obligation. General information only until we understand your situation.
Is Outcome-Based Pricing the Future of SaaS?
For some categories, yes. For most, it'll be one meter among several.
Outcome pricing fits where software replaces a well-defined unit of human work that's already counted and costed: support tickets, sales qualification, recruiting, claims processing, collections. Expect it to spread there.
It fits badly where value is diffuse (a design tool, a project tracker, a CRM), where results take months, or where the customer's own team does most of the work. In those categories, seats with AI credits, or a platform fee plus usage, will stay the norm.
My prediction for the next few years? The winners will offer an outcome price as the headline, a committed volume for the finance team, and a usage meter for anything that doesn't fit either. Salesforce is already there, with every meter at once. Smaller companies will do it with two.
What to Do Next
- List the outcomes your product or service produces, and run each through the four tests.
- Pick one outcome that passes all four. If none do, use usage or credits instead (see the Pricing Model Picker).
- Measure your success rate and cost per attempt on real work for at least a month.
- Model revenue and margin at your current success rate and at 20 points lower.
- Write the definition using the checklist, and test it on two friendly customers.
- Add a floor: a platform fee, a retainer or a minimum commitment.
- Publish it plainly. Fin's pricing page explains exactly what counts. Yours should too.
For real-world inspiration across support, sales, recruiting and services, read outcome-based pricing examples. If you're pricing an AI agent specifically, the guide to AI agent pricing covers all six models with current prices.
Read the whole book free
The Subscription Playbook as a free ebook and audiobook. Australian business owners can claim a free hardcover, shipped free.
Claim the free hardcover Get the ebookMore in this section
Frequently asked questions
Do customers actually prefer outcome-based pricing?
They like hearing it, but many don't choose it. Decagon offers per-resolution and per-conversation pricing and says the vast majority of its customers choose per-conversation. Finance teams want a bill they can forecast. Lead with the outcome story, then let buyers convert it into a committed volume if they prefer predictability.
What happens when the outcome depends partly on the customer?
Write it into the definition. If a hire fails because the client changed the role, or a lead goes cold because their sales team never called, you shouldn't carry that risk. Exclude outcomes the customer controls, set a time window, and agree in advance what evidence counts. Paraform reportedly uses a 90-day replacement guarantee to handle this for hires.
How do I forecast revenue with outcome-based pricing?
Forecast three numbers separately: the customer's volume (tickets, leads, roles), your success rate, and your price per outcome. Multiply them for each account. Track the success rate weekly, because it moves faster than the other two. Minimum commitments give you a floor, so forecast committed revenue and outcome revenue above the commitment separately.
How is outcome-based revenue recognized in the accounts?
Outcome fees are usually treated as variable consideration under revenue standards such as AASB 15 in Australia (the local equivalent of IFRS 15). In simple terms, you recognize revenue when it's highly probable it won't reverse, which can mean later than you'd like. Get your accountant to review the contract before you sign the first big deal.
Can a customer hide outcomes to avoid paying?
If outcomes are measured in the customer's system, yes, which is why most vendors measure them inside their own product. Fin decides an outcome from the conversation itself. For results measured elsewhere, such as revenue or hires, build in audit rights, data access and a fallback fee if reporting stops.
Is outcome-based pricing the same as a success fee?
A success fee is an older, simpler version: a one-off payment when a deal closes, a hire starts or a case is won. Outcome-based pricing in software applies the same idea to many small, repeated results, like each resolved ticket, usually billed monthly. The design questions are identical: define the outcome, measure it, and share the risk fairly.
Get both books free
The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.
Robert Coorey, MBA
Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.
Sources
- Intercom / Fin pricing (checked 27 Oct 2026)
- Intercom Help, Fin AI Agent outcomes
- Zendesk pricing (checked 27 Oct 2026)
- Zendesk, Outcome-based pricing
- Decagon, Pricing AI agents
- Pricing Innovation, Agentforce's pricing journey (6 Oct 2026)
- EngagePulse, HubSpot Breeze pricing and credits explained (17 Jul 2026)
- Value Add VC, How Sierra makes money (updated 21 Oct 2026)
- Eoghan McCabe, There is exactly one way that SaaS can be saved (Fin Ideas, 2025)
- SaaStr summary of ICONIQ 2026 State of AI (10 Jul 2026)
Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.