Contract Mechanics for a Solo FDE: Stage the Risk Down a Ramp — Fixed-Fee Pilot → Milestone-Staged Retainer, with Outcome-Based Kept as Optional Upside, Never the Backbone
The question
Verbatim: "What contract structures (outcome-based, retainer, milestone) let a solo forward-deployed AI engineer de-risk the buyer while protecting margin, and how do buyers perceive solo vs. studio for agent-deployment work?" Context: the FDE cluster has settled the positioning (solo-as-feature) and the price band ($15K/mo retainer); this brief fills the remaining gap — the actual contract plumbing underneath the offer. Deepens [[2026-05-30-solo-vs-studio-fde-buyer-perception]] and [[2026-05-31-fde-scoping-pricing-vs-ai-consultant-framing]].
What we already know (from the vault)
- The price shape is decided; the contract plumbing was not. Retainer-primary at ~$15K/mo (range $12K-$18K core, $25K-$30K intensive ceiling, $7K-$9K light-touch floor), 90-day minimum, no day-rate, no hourly billing, optional small equity kicker. This brief adds the payment/risk mechanics that sit under that number. [[2026-06-02-fde-retainer-band-pricing]], [[2026-05-31-fde-scoping-pricing-vs-ai-consultant-framing]]
- The first-touch offer is already a de-risking instrument: a paid fixed-scope pilot that credits 100% into the retainer, plus a fixed-scope money-back guarantee. Paid pilots convert 40-60% vs <10% for free; predefined success criteria drive a 3.2x conversion lift; the credit-the-fee mechanic (10-30% of contract value, fully credited on conversion) removes the price objection while keeping skin in the game. [[2026-06-02-first-touch-offer-data-buyer]]
- The four named bus-factor neutralizers are contractual, not rhetorical: (1) handoff-ready documentation as a standing deliverable, (2) a named backup/network bench, (3) milestone/staged scoping that caps buyer exposure at any single point, (4) a small paid pilot. Contract structure is how a solo manufactures the "perceived safety" a studio gets from its bench. [[2026-05-30-solo-vs-studio-fde-buyer-perception]]
- Solo reads as a feature for a build-shaped engagement and a liability for an open-ended seat. The contract must therefore stay time-boxed and deliverable-anchored (≤90 days to production, hand-back) — the moment it drifts to "be our head of data indefinitely," the solo bus-factor read strengthens and the studio's bench argument gets teeth. [[2026-05-30-solo-vs-studio-fde-buyer-perception]]
- "The artifact IS the deliverable" — tracking billable hours is the consulting-trap warning sign. Any contract structure that meters hours fights RDCO's own model; the structure must price a thing shipped over a bounded window. [[2026-05-31-fde-scoping-pricing-vs-ai-consultant-framing]], [[2026-05-13-fde-asymmetric-edge-rdco-positioning]]
What the web says
- The four models map cleanly to a risk-allocation gradient. Hourly: client bears scope/duration risk (mitigated by a weekly cap — "we don't run open-ended hourly work without a weekly cap"). Fixed-fee project ($10K-$75K): the consultant carries delivery risk, protected by tight scope + change-order clause. Retainer ($5K-$15K SMB, $15K-$25K enterprise, 6-12mo): consultant carries underutilization risk, protected by a minimum monthly revenue floor. Outcome fee: consultant bears all execution + attribution risk. (Dojo Labs)
- The market's own answer is a hybrid 3-phase ramp, not a single model. Dojo Labs structures engagements as scoping ($3,500 / 2 weeks, fixed) → build ($15K-$40K / 8-12 weeks, project) → maintenance ($6K-$10K/mo, retainer) — "low-risk discovery upfront, then fixed-price execution, then ongoing." McKinsey's QuantumBlack runs the same shape: a fixed base engagement fee + tiered resource packages + success fees tied to outcomes. (Dojo Labs, Consulting Success)
- Milestone staging is the core de-risk-both-sides device, with standard parameters. 30-50% upfront to cover discovery/setup, remainder split across milestones (templates: 40% kickoff / 20% design / 20% dev / 20% go-live). Each milestone needs a written acceptance bar ("UX approved," "MVP deployed"), a review window (commonly 5 business days, with silence = auto-accept), a change-order clause (new asks trigger revised cost/time), and a kill fee / cancellation clause defining what % is payable on termination-without-cause. Shorter milestones (≤2-3 weeks) give the provider steady cash flow while the buyer only pays after value lands. (Skydo, DealHub, fynk)
- Outcome-based pricing is real but is almost always a hybrid, and it hands the vendor the attribution problem. Typical structure: 10-25% of verified gains, or a reduced base (often ~50% lower upfront) plus a success bonus, released only after the metric is confirmed over a 30-90 day window. Sellers manage the risk with "caps, minimums, or hybrid structures." Sierra's agent model charges per resolved outcome ("if the conversation is unresolved, in most cases there's no charge") but falls back to consumption-based pricing when outcomes don't fit — i.e., even the poster child hedges. (Dojo Labs, Sierra, Zendesk)
- Value-based pricing prices a fraction (typically 10-25%) of the quantified outcome — lucrative when attribution is clean, but it "requires solid baseline metrics to measure against" and only works for "revenue-critical, measurable workflows." For data-plumbing/agent work, the outcome is often a step removed from revenue, so attribution is contestable. (Consulting Success, Dojo Labs)
- Buyers de-risk small/solo vendors primarily through milestone escrow and evidence-based acceptance. Funds held by a neutral third party (Escrow.com, Castler) release per milestone only after a defined inspection window (3-15 days typical) and evidence requirements; disputes go to a mediation step before release. Source-code/IP escrow is also cited as a continuity hedge "when the vendor fails." These are exactly the devices a solo can offer proactively to flip the bus-factor script. (Skydo, Escrow.com, ITPro)
Convergences and contradictions
- Strong convergence (web ↔ vault): nobody sells a single pure model — the winning shape is a staged ramp from fixed-fee diagnostic to retainer. The web's independent answer (Dojo Labs 3-phase, QuantumBlack base+success) is the same instrument the vault already assembled piecemeal (paid pilot → 90-day retainer). The two literatures converge on "stage the risk down, don't pick one model."
- Convergence: milestone staging is the device that de-risks the buyer AND protects vendor cash flow simultaneously — the buyer pays only after value lands; the vendor gets steady cash and a fixed scope per stage. It is the one mechanism both sides of the table want, which is why it is the natural backbone for a solo who needs the buyer's trust and can't float a long unpaid build.
- Genuine tension to manage (and it cuts harder for a solo): the web markets outcome-based pricing hard ("2.3x client satisfaction," McKinsey/Sierra showcases) but every source quietly hedges with base fees, floors, caps, and consumption fallbacks. "Pure" outcome-based barely exists in practice. For a studio with a bench, eating a missed-outcome quarter is a margin dent; for a solo with no bench to absorb the miss and real cash-flow sensitivity, it is closer to existential. The attribution problem (was it the agent, or the client's market?) lands entirely on the vendor and is contestable for data/agent work where the outcome is a step removed from revenue. So the web's enthusiasm for outcome-based should be down-weighted for RDCO's structure, not adopted at face value.
- No contradiction on the solo-vs-studio read — and the contract mechanics resolve it. The vault's finding (solo = feature for a build, bus-factor = the one real objection) plus the web's de-risk toolkit (milestone escrow, acceptance windows, code+docs handoff, kill fee) means a solo can contractually manufacture the perceived safety a studio gets from headcount. The studio buys safety with a bench; the solo buys it with structure — and structure is cheaper and reads as confidence, not apology.
Synthesis for RDCO
Recommended contract shape: a two-stage risk ramp where each stage de-risks the buyer further while the structure — not a contingency bet — protects margin. Stage 1 is the already-decided paid fixed-fee pilot ($2K-$5K, ≤2-3 weeks), governed by a one-page SOW with a written, agreed success bar, a fixed-scope money-back guarantee ("delivered to the agreed spec by the agreed date, or refunded"), and 100% fee-credit into the retainer on conversion. This is the buyer's headline de-risk: small dollars, short window, guaranteed, and free to proceed. It is cheap for RDCO to guarantee precisely because the scope is tiny and the window is short — the guarantee is a high-trust signal a solo can afford that a studio rarely matches at the individual level. Stage 2 converts to the 90-day-minimum retainer at $15K/mo, but billed as milestone-staged releases inside the term rather than one upfront lump: month-one (or a kickoff deposit) up front to seat the engagement, then releases tied to shipped production artifacts, each with a 5-business-day acceptance window (silence = auto-accept), documentation + working-code handoff as a standing line-item deliverable, a change-order clause so every new ask re-prices scope and time, and a kill-fee/cancellation clause (e.g., 30 days' notice or pay-for-work-completed-to-date). The retainer floor is the margin backbone; the milestone staging is the buyer's exposure cap; the change order is the scope-creep firewall.
On outcome-based: keep it as optional, capped upside layered on top of a guaranteed floor — never the backbone, and never pure. The evidence is clear that pure contingency hands a solo all the execution and attribution risk with no bench to absorb a miss and a cash-flow profile that can't survive a 30-90 day unpaid verification window. If a buyer pushes for skin-in-the-game beyond the guarantee, offer a hybrid: the full $15K/mo retainer floor stays, plus a small success kicker (10-25% of a single, pre-agreed, attributable metric the buyer signs off on as caused by the build), capped at a defined ceiling, and only where a clean baseline exists. For most first data/agent engagements the outcome is a step removed from revenue (a reliable pipeline, a working eval harness, a deployed agent), so attribution is contestable and the kicker should be the exception, not the default. The money-back guarantee on the pilot already gives the buyer the de-risk that outcome-based promises, without exposing solo margin to a metric RDCO doesn't fully control.
The solo-vs-studio answer, in contract terms. Buyers read a solo as senior-does-the-actual-work (the studio's biggest grievance, "senior pitch, junior delivery," is structurally impossible solo) — a feature for a deep single-discipline build. The lone liability is bus-factor/continuity, and every neutralizer is a contract clause RDCO offers unprompted: milestone staging caps exposure at any single point; the money-back guarantee caps downside on the trial; code-plus-documentation handoff as a standing deliverable means the buyer is never stranded; and for a genuinely nervous first enterprise buyer, offer third-party milestone escrow proactively (Escrow.com-style, funds release per accepted milestone). Offering these before they are asked for flips the script: the solo stops being the risky default that needs excusing and becomes the operator confident enough to put the downside in writing. The studio buys safety with a bench the buyer pays for; RDCO buys it with structure the buyer keeps.
Net contract recipe: paid fixed-fee pilot (guaranteed, credited) → 90-day-min retainer at $15K/mo, milestone-staged with 5-day acceptance windows, change-order + kill-fee clauses, code+docs handoff as a deliverable, escrow offered on request, and outcome-based confined to an optional capped kicker on a pre-agreed attributable metric. Keep it build-shaped; the whole structure degrades the moment it becomes an open-ended seat.
Open follow-ups
- Draft the actual one-page SOW + lightweight MSA template encoding the milestone schedule, acceptance window, change-order, kill-fee, guarantee, and handoff language — then run it past a contracts attorney before the first real pitch (the guarantee + refund terms especially).
- Set the parameters that are still placeholders: kill-fee % (30-day notice vs pay-to-date), acceptance-window length (5 business days is the web default), and the milestone split for a 90-day build (e.g., kickoff deposit + 2-3 artifact-gated releases).
- Decide the escrow posture: offer milestone escrow as a standing option or only when a buyer asks? Pick a provider (Escrow.com / Castler) and pre-write the clause so it's frictionless to add.
- Decide the success-kicker policy: offer it at all? If so, which metrics on a data/agent deliverable are clean enough to be attributable, what's the cap, and does it conflict with the phData W-2 single-client-concentration guardrail (same constraint flagged on the equity kicker).
- Get one real engagement to reveal what buyers actually counter with — milestone staging, full upfront, or escrow — converting this analogical contract design into RDCO-specific data.
Related
- [[2026-05-30-solo-vs-studio-fde-buyer-perception]] — the solo-as-feature finding and the four bus-factor neutralizers this brief converts into contract clauses.
- [[2026-05-31-fde-scoping-pricing-vs-ai-consultant-framing]] — "the artifact IS the deliverable," no-billable-hours, retainer-shape decision the milestone staging sits inside.
- [[2026-06-02-first-touch-offer-data-buyer]] — the paid pilot + fee-credit + money-back guarantee that is Stage 1 of the ramp.
- [[2026-06-02-fde-retainer-band-pricing]] — the $15K/mo retainer band the Stage 2 milestone schedule prices against.
- [[2026-06-03-fde-retainer-monthly-delivery-model]] — what ships each month to justify the retainer the milestones gate.
- [[2026-05-23-agent-deployer-competitor-pricing-scan]] — the three-tier price map and the saturated audit tier to avoid.
- [[2026-05-13-fde-asymmetric-edge-rdco-positioning]] — the $5K-$30K artifact-and-template band these contracts operate within.
Sources
Vault:
- ~/rdco-vault/06-reference/research/2026-05-30-solo-vs-studio-fde-buyer-perception.md (solo-as-feature; four bus-factor neutralizers; build-shaped guardrail)
- ~/rdco-vault/06-reference/research/2026-05-31-fde-scoping-pricing-vs-ai-consultant-framing.md (artifact-as-deliverable; no-hours; retainer shape; vendor hourly $90-$300/hr)
- ~/rdco-vault/06-reference/research/2026-06-02-first-touch-offer-data-buyer.md (paid pilot 40-60% conversion; fee-credit 10-30% of ACV; money-back guarantee; predefined success criteria 3.2x lift)
- ~/rdco-vault/06-reference/research/2026-06-02-fde-retainer-band-pricing.md ($15K/mo anchor; $12-18K core; $25-30K ceiling; $7-9K floor; $2-5K one-time pilot)
- ~/rdco-vault/06-reference/research/2026-06-03-fde-retainer-monthly-delivery-model.md (monthly delivery cadence the milestones gate)
- ~/rdco-vault/06-reference/research/2026-05-23-agent-deployer-competitor-pricing-scan.md (three pricing tiers; audit tier saturated)
- ~/rdco-vault/06-reference/concepts/2026-05-13-fde-asymmetric-edge-rdco-positioning.md ($5K-$30K band; $30K phData hand-up ceiling)
Web (accessed 2026-06-07):
- Dojo Labs — AI Consulting Pricing Models Explained (4-model risk gradient; 3-phase hybrid ramp $3.5K scope / $15-40K build / $6-10K/mo; outcome 10-25% of verified gains, 30-90 day verification): https://dojolabs.co/blog/ai-consulting-pricing-models/
- Sierra — Outcome-Based Pricing for AI Agents (pay-per-resolved-outcome; consumption fallback when outcomes don't fit): https://sierra.ai/blog/outcome-based-pricing-for-ai-agents
- Skydo — Milestone Payments: How Businesses De-risk Projects (30-50% upfront; 5-business-day review/auto-accept; kill-fee + change-order clauses; ≤2-3wk milestones for cash flow): https://www.skydo.com/blog/milestone-payment-system
- Consulting Success — From Hourly to Value-Based Pricing (10-25% of quantified outcome; QuantumBlack base+resource+success hybrid): https://www.consultingsuccess.com/value-based-pricing-consultants
- Zendesk — Understanding Outcome-Based Pricing (caps, minimums, hybrids to manage seller risk): https://www.zendesk.com/blog/ai/agentic-ai/outcome-based-pricing/
- DealHub — What Is Milestone-Based Pricing (milestone definition, acceptance, release conditions): https://dealhub.io/glossary/milestone-based-pricing/
- fynk — Milestone Payment Schedule clause (acceptance criteria, evidence, review windows): https://fynk.com/en/clauses/milestone-payment-schedule/
- Escrow.com — Milestone & Service Transactions (neutral fund hold; per-milestone inspection + release): https://www.escrow.com/milestones
- ITPro — Why B2B buyers need escrow when the vendor fails (source-code/continuity escrow as solo-vendor hedge): https://www.itpro.com/business/business-strategy/when-the-vendor-fails-why-b2b-buyers-need-escrow-as-a-priority-for-their-software-stack