Results-Based PR in 2026: What the Fee Buys and What It Doesn't
Results-based PR ties the fee to a defined result. AuthorityTech guarantees the agreed placement, backed by escrow; traffic, AI citations and ROI are outcomes to measure afterwards, not promised.
Disclosure: AuthorityTech publishes this page and is the placement-fee example in it. Its terms below follow its Statement of Work (SOW) and public Refund Policy as the cleared guaranteed-media terms page states them, not as an independent finding. Machine Relations Research, which publishes the Machine Relations Index cited below, and AuthorityTech are both Bramwell companies, Machine Relations Research and AuthorityTech are Bramwell companies, the Index is theirs, and Jaxon Parrott created it.
Results-based PR is a PR model in which the fee is tied to a defined result instead of to a period of activity. AuthorityTech guarantees the agreed placement, backed by escrow and the refund terms in its SOW; whether that placement later brings traffic, leads, AI citations or revenue is a separate question, to be measured after the fact and not promised in advance. AuthorityTech's SOW works this way: the fee sits in escrow, is released when publication is confirmed, and is refunded in full if the agreed placement is not delivered, subject to the SOW's exclusions. It has no traffic, citation or ROI term. In Machine Relations terms, the placement is the delivered input to an AI-search system, and the downstream outcomes are what you measure against a dated baseline.
Key Takeaways
- The guaranteed result is the agreed placement. AuthorityTech's SOW commits to an agreed placement; the publication decides what runs.
- Escrow moves delivery risk, not outcome risk. If the agreed placement is not delivered the escrow is refunded; if it is delivered and moves nothing downstream, the client still carries that risk.
- Downstream outcomes are measurable but not owed. Referral traffic, leads, AI mentions and citations can be recorded with a baseline and a window; none of them is a term of the SOW.
- Engines cite a mix of sources. In the Machine Relations Index, editorial publications take the largest share of role-classified buyer-question citations, with company websites close behind.
- Ask for the terms in writing. Judge any results-based offer on what it defines as the result, who decides, and what triggers a refund.
What results-based PR means, and what it cannot mean
"Results-based" and "performance" PR are used loosely for pay-for-results and pay-per-placement services, and the label does not say what the result is. The useful question is what the contract defines as the result that triggers payment. In a placement-fee model it is a published placement in an agreed outlet. It is not a promise that readers, engines or buyers will respond.
Machine Relations Research separates the two in its Performance PR definition: it uses the term as a measurement convention for evaluating coverage against declared, observable post-publication outcomes, and keeps causal explanations separate from the observations. A placement is recorded once; outcomes are recorded afterwards, against a window and a baseline, with the limits written down. That is the difference between what a contract delivers and what a team measures. The AuthorityTech Performance PR glossary entry and the performance PR guide cover the model itself.
What AuthorityTech's contract delivers
AuthorityTech's SOW has the client pay the fee into escrow before securement begins, releases it when publication is confirmed, and refunds it in full if the agreed placement is not delivered, subject to the SOW's exclusions. The SOW states that the publication decides what runs and has no citation, traffic or ROI term. The public Refund Policy separately sets a 90-day guarantee with refund of the escrow or a replacement placement of equal or greater value, on conditions such as timely client materials and approvals; outside that guarantee and the SOW's own terms it treats payments as final. AuthorityTech sets its fee per engagement, publishes no rate card, and engagements are invite-only, per its engagement page. That page describes the model as pay-after-publication with funds held in escrow until a placement is live; the SOW, which controls under AuthorityTech's Terms, has the fee paid into escrow before securement and released on publication. Read the contract you are offered, since both descriptions agree that money is held until the placement is live.
What is hoped for, not promised
These are outcomes a team might use to judge whether a placement worked. None of them is a term of AuthorityTech's SOW, and each has its own way of being measured:
| Outcome | In AuthorityTech's SOW? | How a team can measure it |
|---|---|---|
| An agreed placement is published | Yes: the delivery term, with an escrow refund if it is not delivered | The live URL, outlet and date |
| What the piece says and whether it runs | No: the publication decides | Read the published piece against the brief |
| Referral traffic and leads | No | Your own analytics and CRM, tied to the placement URL |
| Revenue or ROI | No | Your own pipeline data; attribution is a judgment, not a given |
| AI mentions and citations | No | A dated baseline on a fixed question set: brand mentions, cited hosts and exact URLs per engine, repeated after a stated window |
Ask any provider, AuthorityTech included, what reporting you receive, in what form and how often, and get it in writing; AuthorityTech's public Terms mention reporting dashboards, and its SOW is the document that controls.
What the evidence says about whether a placement gets cited
In the Machine Relations Index release of November 1, 2026, across the 131 published strata for its six buyer question shapes (best-of, how-to-choose, is-it-worth-it, problem-first, top-list and X-versus-Y), editorial publications took 33.3% of the citations that went to a role-classified domain, vendor-owned sources (company websites) 27.8%, community and social platforms 12.6%, academic and government sources 11.8% and wire and press-release distribution 0.8%. Vendor-owned sources were ahead of editorial publications on how-to-choose questions (30.4% to 29.7%) and on problem-first questions (26.4% to 26.2%). Those shares rest on the 27.8% of buyer-question citations that land on a role-classified domain.
Those shares describe which sources were observed in answers for the Index's monitored prompts, not what any placement produces, and a company's own website is classed as vendor-owned whoever the company is. In the same release, forbes.com was cited in 1,198 of 34,788 monitored answer runs, techcrunch.com in 177 and wsj.com in 130, so which outlet an engine cites differs by nearly an order of magnitude even among well-known business titles. that spread is a reason to check which outlets engines cite for your own category and question before agreeing to a placement, rather than to assume one outlet tier works for every question.
Muck Rack's October 2026 What Is AI Reading? edition found that more than 84% of cited links came from earned and third-party sources, meaning sources brands neither own nor pay for (journalism, government and NGO websites, academic research, encyclopedic sources and third-party corporate content). In September owned media was 15.2% of citations and paid and advertorial content 0.3%, and journalism alone was 22%, down from 27% in May.
None of this shows that a given placement will be cited. It shows what kinds of sources engines draw on, and it is why a placement is worth measuring rather than assuming.
Results-based versus retainer: what each contract settles
BuzzStream's digital PR cost survey reports an average monthly contract of $5,458 across all respondents, about 70 digital PR agency leaders and freelancers, with most retainers under $10K a month and about half below $5K, and an average cost per link of $597. That survey covers about 70 digital PR agency leaders and freelancers, not AI PR and not placement-fee models. See also the PR retainer trap for the argument against retainers.
| Question to settle | Retainer agency | Placement-fee agency such as AuthorityTech |
|---|---|---|
| What the fee is for | A scope of work over a period, as the agreement sets it out | An agreed placement |
| What happens if the deliverable does not appear | Whatever the agreement says; ask before signing | AuthorityTech's SOW refunds the escrow in full, subject to its exclusions |
| Who decides what runs | The publication, in either model | The publication, in either model |
| Does the contract promise traffic, citations or ROI | Read the agreement | No: AuthorityTech's SOW has no such term |
How to evaluate any results-based PR offer
- What exactly counts as the result? A published placement, a mention, a link, a broadcast segment? Does syndication or a sponsored piece count?
- Who decides what runs, and when is the result "delivered"? Look for the confirmation trigger and the timeline.
- What triggers a refund, and what are the exclusions? A refund for a missed placement is not a refund for a placement that did not move your numbers.
- What is reported afterwards? Ask for the format and cadence in writing, not "tracking".
- How is the fee set, and when is it disclosed? Fees set per engagement mean you need the number before comparing.
Measuring what a placement does, the Machine Relations way
To measure downstream outcomes rather than assume them, record the placement (URL, outlet, date), take a dated baseline on a fixed question set before publication, choose an observation window, and then record mentions, cited hosts, exact URLs and referral visits at the end of it, with the limits written down. Treat any change as a lead, not proof of cause: the engines, the question, freshness and competing sources can all affect what is cited. The startup PR guide sets out the same baseline method in more detail, and Machine Relations Research's Performance PR record is the template for writing the observation down.
GEO and AEO are distribution practices inside the broader Machine Relations system, and the placement is the earned-authority input they depend on. The Machine Relations stack describes each of its layers as a hypothesis to test, which is the right posture for a placement too. AuthorityTech offers a visibility audit that can help take a baseline; confirm that whichever tool you use records mentions, cited hosts and exact URLs per engine.
Frequently Asked Questions
What is results-based PR?
Results-based PR is a model in which the fee is tied to a defined result, usually a published placement, instead of to a period of activity. The label does not say what the result is, so read the contract for the definition, the trigger and the refund terms.
Does results-based PR guarantee citations, traffic or ROI?
Not in AuthorityTech's SOW. It commits to an agreed placement with an escrow refund if the placement is not delivered, subject to exclusions, and has no citation, traffic or ROI term. Those outcomes can be measured afterwards against a baseline, but a contract that describes them as guaranteed deserves a close read.
What does AuthorityTech's SOW deliver?
An agreed placement. The client pays the fee into escrow before securement begins, it is released when publication is confirmed, and it is refunded in full if the agreed placement is not delivered, subject to the SOW's exclusions. The publication decides what runs. The public Refund Policy also sets a 90-day guarantee with refund or replacement, on conditions.
How is a placement-fee model different from a retainer?
A retainer is a fee for a scope of work over a period, as the agreement sets it out. A placement-fee model ties the fee to an agreed placement. Neither controls what a publication runs, and neither makes citations or ROI a certainty; compare the terms you are offered.
How do I tell whether a placement worked?
Decide the outcomes before publication, take a dated baseline on a fixed question set, and compare mentions, cited hosts, exact URLs, referral visits and pipeline after a stated window. Treat the comparison as a lead rather than proof, because the engines, the question and competing sources also change what is cited.
Which outlets get cited by AI engines?
It varies by outlet and by question. The Machine Relations Index publishes citation counts for each domain, so you can look up an outlet and a category before choosing it.
Start with the AuthorityTech visibility audit or your own baseline, then decide which placement to agree. For the terms themselves, see the guaranteed-media terms page.