Build an Experience Marketplace: The trust.guide Playbook

The problem

The creator economy hit $254 billion in 2025 and is projected to clear $320 billion by 2026. Underneath the headline number sits a structural problem no major platform has solved: anyone can call themselves an expert, and buyers have no way to verify the claim before paying. Udemy hosts over 200,000 courses with no mechanism to validate that the instructors have actually done what they teach. Substack writers monetize newsletters on topics they may have only researched, never practiced. Up to 14% of online reviews are likely fraudulent — more than 2.3 million generated by AI in 2024. Consumer trust in online reviews plunged from 76% in 2019 to 46% by 2022, and the curve has steepened since.

When someone faces a genuinely unfamiliar situation — relocating to a new country, navigating a serious medical diagnosis, starting a business in a foreign jurisdiction — they do not need more content. They need borrowed experience from someone who has actually done it, tested by someone who is about to do it, and structured so they can act on it immediately. The creator economy has a verification gap, not a content gap.

This walks through the three-party architecture that closes the verification gap: contributor, validator, platform, each with different skills and different economic incentives. The KYB pattern for expert verification, the composite trust score that replaces star ratings, the 15-25% tiered commission model, the AI-driven content extraction pipeline, and the 30-day launch playbook for solving the cold-start supply problem.

What most people get wrong

Mistake one: requiring experts to also be content creators. Every major content platform shares this flaw — the person who creates the content is also the person who claims it works. Medium writers assert expertise through prose quality. Udemy instructors demonstrate knowledge through lecture delivery. The trust signal is the creator’s own presentation, not independent verification of their claims. The result is a market where the loudest voice wins, not the most experienced. SEO-optimized articles outrank practitioners who have never written a blog post. The people with the most valuable experience — the ones who successfully relocated, started a business, navigated a legal system — are often the worst at packaging that experience for sale, because their skill is in doing, not in marketing. Decouple the roles. The expert needs one skill: the ability to describe what they did. The platform handles AI-driven extraction, structuring, distribution.

Mistake two: rating-based trust instead of validation-based trust. Star ratings, upvote counts, and testimonial carousels are noise, not signal. The UK’s Digital Markets, Competition and Consumer Act began enforcement in July 2025; it reviewed 100 sites and found 54 in breach of review authenticity requirements. Any marketplace that relies on unverified star ratings inherits that credibility deficit. Replace ratings with validation: real people, in real situations, documenting outcomes. A guide with three independent validators who followed the steps in their own situation and confirmed the outcomes carries a fundamentally different trust signal than a guide with five hundred star ratings that may include 14% AI-generated noise. The structural innovation is to make validation a paid role in the marketplace, not a side effect of customer behavior.

A working approach

The three-party model. Three roles, three different skills, three different economic positions:

Role	What they do	Skill required	Revenue share
Contributor	Shares validated experience via interview	Domain experience	50%
Validator	Tests the guide in a real situation	Currently doing it	10%
Platform	Extracts, structures, sells, distributes	Curation + tech	40%

The contributor needs domain experience, not writing skill. A 45-year-old entrepreneur who started three companies in Europe but has never written a blog post can participate fully — the platform extracts her knowledge through a 60-minute structured conversation, AI-structures it into a sellable guide, and she earns 50% of every sale, indefinitely. The validator solves the cold-start trust problem: a new guide with zero sales history can demonstrate credibility if three validators independently confirm the steps work, with timestamped evidence. The platform is the curator, structurer, marketplace operator — explicitly not the content creator, which avoids the liability of claiming expertise while capturing the value of organizing and distributing verified knowledge.

KYB for expertise. Three tiers of expert verification, each providing a progressively stronger trust signal:

interface ExpertVerification {
  verificationRules: {
    "business-formation": {
      required: ["business_registration", "tax_id"],
      optional: ["revenue_proof", "employee_count"],
      expiry: "12_months"
    },
    "legal-navigation": {
      required: ["bar_admission", "jurisdiction"],
      optional: ["years_practicing", "specialization"],
      expiry: "24_months"
    },
    "relocation": {
      required: ["destination_residency_proof"],
      optional: ["years_in_country", "language_certification"],
      expiry: "36_months"
    }
  };
}

Tier 1 (self-declared) lowers the barrier — anyone can start contributing immediately, no badge, lower search ranking. Tier 2 (document-verified) takes professional licenses, business registration, certifications, runs document extraction via OCR and cross-reference against public registries; Silver badge. Tier 3 (community-verified) requires three or more successful guide validations and peer endorsements; Gold badge. Financial institutions running automated KYB report 65% faster onboarding and 40% fewer false positives versus manual; the same automation principles apply to expert verification.

The composite trust score replaces star ratings. Five inputs, each weighted:

interface TrustScoreInputs {
  validationOutcomes: number;      // 35% — did validators confirm?
  buyerSatisfaction: number;       // 20% — post-purchase weighted
  credentialVerification: number;  // 20% — KYB tier
  responseQuality: number;         // 15% — Q&A and consultation
  longitudinalAccuracy: number;    // 10% — does the guide age well?
}

function computeTrustScore(i: TrustScoreInputs): number {
  return Math.round(
    i.validationOutcomes * 0.35 +
    i.buyerSatisfaction * 0.20 +
    i.credentialVerification * 0.20 +
    i.responseQuality * 0.15 +
    i.longitudinalAccuracy * 0.10
  );
}

Anti-gaming is structural: validators must be independent users (not contributor-controlled accounts), validation outcomes are time-stamped and require documented evidence, credential verification runs through external registries the contributor does not control, and longitudinal accuracy comes from buyers who return after acting on the guide — which contributors cannot fake. The 5-level reputation hierarchy (Apprentice through Master) is derived from the composite score and surfaces directly in search ranking.

Commission architecture. Tiered 15-25% rather than flat:

Volume tier	Platform commission	Contributor net
0-€10K lifetime	25%	50% (of remainder after 10% validator)
€10K-€100K	20%	55% (with 10% validator)
€100K+	15%	60% (with 10% validator)

Commission calculation is server-side only — never trust client-sent amounts. The volume tier is computed from lifetime gross revenue at transaction time, the commission split is applied to the net-of-payment-processor amount, and the audit trail records the calculation inputs alongside the output so any future dispute resolves against immutable evidence.

Where this scales

The article above is the spine. The full book covers what survives the launch and the first thousand guides:

• The content extraction pipeline — Stage 1: structured live interview protocol (60 minutes, scripted to extract the steps, costs, traps, resources). Stage 2: AI-powered extraction that turns transcript into structured guide draft. Stage 3: quality structuring with the section template (Background, Prerequisites, Steps, Costs, Traps, Resources, FAQ). Stage 4: contributor review for accuracy. Stage 5: validator-driven testing in a real situation.
• Consultation and workshop commissions — pricing tiers from Guide Only (€19-49) through Guide + Q&A (+€30), Guide + Consultation (+€100), Workshop (€199-499), Mentorship (€299-999/month). Different unit economics per tier; the book walks through the unit-economic math.
• Professional directory revenue — the second-order monetization layer that surfaces verified experts to enterprise buyers who need consulting or fractional engagements. Higher AOV, longer cycle, different commission curve.
• Ecosystem integration patterns — the wish-to-guide pipeline (when a user on wish.now says “I wish someone could explain how to start a company in Poland” the system routes that wish to a verified trust.guide contributor whose guide has been validated by real people). Identity and payment sharing across the nine-platform mesh.
• The 30-day launch playbook — Week 1: niche selection and first outreach (the cold-start problem requires choosing one niche, not three). Week 2: first guide extraction (live interview protocol). Week 3: validation cycle (three validators, real situations). Week 4: launch and first sales. Outreach templates, success metrics per phase, and the failure-mode catalog for what can go wrong.

Included with the book

• guide-creation-template.md + PDF — structured interview script the platform uses to extract experience from contributors who are not writers. Section-by-section template for the guide format (Background, Prerequisites, Steps with sub-steps, Costs with currency, Traps, Resources, FAQ). Drop in as the source-of-truth template for your own content pipeline.

Get the full picture

Readers of this also chose

Questions readers ask

Is this a clone of trust.guide or a playbook I can adapt to my own niche?

The book is the architecture and operational playbook — drawn from the trust.guide implementation but written so you can apply it to any expertise vertical. The three-party model, the KYB tiers, the trust score formula, the commission structure, the launch playbook all generalize. The case studies cite trust.guide specifics; the implementation patterns are not vertical-locked.

How does the validator role actually scale? Three validators per guide sounds expensive.

Three validators per guide is the launch bar; mature marketplaces converge on one or two for established contributors with high trust scores. The validator economics work because the validator earns 10% of all future guide sales — a successful guide that sells for years compounds the validator's return. The book walks through the math: a €39 guide that sells 200 copies per year over three years pays a validator €234, which is well above the time-cost of one real-situation walkthrough.

What about regulated fields — legal, medical, financial?

The disclaimer architecture chapter walks through the boundary. Regulated fields require Tier-2 document verification (bar admission for legal, jurisdiction-specific license for medical, FCA/SEC equivalent for financial), explicit non-advice disclaimers in the guide template, and a separate insurance line. The book covers the disclaimer copy that satisfies most jurisdictions plus the routing logic that surfaces regulated guides only to buyers in the matching jurisdiction.

How do you prevent contributors from gaming the validator role?

Anti-gaming is structural rather than after-the-fact moderation. Validators must be independent users with their own KYB and purchase history; contributors cannot create validator accounts. Validation outcomes are time-stamped and require documented evidence (screenshots, receipts, status emails). The longitudinal accuracy input to the trust score comes from buyers who return after acting on the guide — which the contributor cannot fake. The combined design closes the most common gaming surfaces.

What's the refund policy?

Lemon Squeezy's standard refund window applies. The refund link is in the receipt email.