Research brief: trust, referral networks, and in-group reputation in Ontario's trades economy (May 2026 — Foundation Brief #3)

Status: Third foundation brief in the psychology cluster, companion to [[research-brief-psychology-gc-marketing-aversion-may-2026]] (Brief #1) and Research brief: risk aversion, loss aversion, and post-failure decision patterns in GC and trades-business decision-makers (May 2026) (Brief #2). May 2026, ~10,500 words. Internal reference; not customer-facing.

TL;DR

• Referred customers in service businesses have ~16% higher LTV than matched non-referred customers. Schmitt-Skiera-Van den Bulte 2011 (10,000 German retail-bank customers, 33 months). Retention advantage persists; margin advantage decays. See [[schmitt-skiera-van-den-bulte-2011-16pct-ltv-premium]].
• WOM elasticity is ~0.53 — roughly 20× marketing-events elasticity, 30× media-appearances elasticity (Trusov-Bucklin-Pauwels 2009, online social network). See [[trusov-bucklin-pauwels-2009-wom-elasticity-053]].
• Trust mediates the strong-tie advantage in useful knowledge transfer. Once competence- and benevolence-based trust are controlled for, the structural benefit of weak ties re-emerges — Levin & Cross 2004. See Levin & Cross 2004 (Mgmt Sci) — competence- and benevolence-based trust mediate tie-strength → useful knowledge transfer; once trust is controlled for, weak ties re-emerge as valuable.
• Closeness is the best tie-strength indicator, NOT frequency or duration of contact (Marsden-Campbell 1984). See Marsden & Campbell 1984 (Social Forces) — closeness is the best tie-strength indicator, NOT frequency or duration.
• The Ontario residential-construction signal stack has three audit-verifiable tiers and two purchasable tiers. HCRA + Tarion + Gold Seal are audit-verifiable; RenoMark is membership-screened-with-attestation; Best of Houzz badges and pay-to-play "Top Contractor" lists are engagement- or fee-driven. See [[signalling-theory-applied-ontario-credential-stack]].
• Guildmaster Awards are audit-verified survey signals, not paid badges; ≥90% Likely-to-Recommend rate against ~70% industry average, ≥20 responses, disqualification for omitting customers. See [[guildmaster-guildquality-audit-verified-survey-signal]].
• In-group reputation and market reputation are not the same construct. Rindova et al. 2005: "being good" (perceived quality among informed stakeholders) vs "being known" (prominence). For high-ticket residential/ICI work, in-group dominates market. See [[rindova-2005-being-good-vs-being-known]].
• The credence-good problem is solved by liability + verifiability, not by reputation alone. Dulleck-Kerschbamer-Sutter 2011 found experimentally that "liability has a crucial, but verifiability at best a minor, effect" on credence-good market efficiency, and "allowing sellers to build up reputation has little influence." See [[dulleck-kerschbamer-sutter-2011-liability-verifiability-credence-goods]].
• Spence's signalling theory predicts purchased signals collapse when cost drops. Best of Houzz badges and pay-to-play lists are low-cost-to-acquire and carry low signalling value among informed GC buyers; HCRA licensure + Gold Seal certification are costly and retain value. See [[spence-1973-job-market-signaling]].
• Peer-coached contractors (BTA / EOS / Vistage / TAB) are a meaningfully different persona. BTA reports 600+ contractor companies across North America; EOS reports 200K+ businesses, 800-1000 implementers; Vistage 45K members. Causal direction (coaching → marketing maturity) unproven; correlation strong enough to act on. See [[peer-coached-contractor-persona-synthesis]].
• The strongest aversion countermeasure available to Candid is becoming a member of the network whose information flows it wants to participate in. Marketing assets without network membership are noise; network membership generates signal whether or not marketing assets exist. This brief's strongest take and a refinement of how to read Brief #1. See Research brief: trust, referral networks, and in-group reputation in Ontario's trades economy (May 2026 — Foundation Brief #3) Section 6 below + [[rule-network-membership-precedes-referral-program-design]].

The eight strategic implications (R1–R8)

1. [[rule-network-membership-precedes-referral-program-design]]
2. [[rule-reweight-marketing-portfolio-toward-in-group-accrual]]
3. [[rule-target-peer-coached-contractor-persona-with-explicit-qualification]]
4. [[rule-earn-one-audit-verified-credential-per-year-refuse-purchased]]
5. [[rule-better-copy-is-wrong-strategy-credence-good-requires-liability]]
6. [[rule-build-candid-audited-reputation-infrastructure]]
7. [[rule-treat-events-as-reputation-witnessing-not-lead-generation]]
8. [[rule-name-credence-good-problem-explicitly-propose-liability-structures]]

Sections

1. Social topology of the Ontario trades network — [[granovetter-1973-strength-of-weak-ties]], Granovetter 1983 (Sociological Theory) — weak-ties claim walked back; strong ties carry relational guarantees weak ties cannot, esp. for trust under uncertainty, Burt 1992 (Structural Holes) / 2004 (AJS) — brokers spanning holes between groups capture information-arbitrage advantage, Marsden & Campbell 1984 (Social Forces) — closeness is the best tie-strength indicator, NOT frequency or duration, Levin & Cross 2004 (Mgmt Sci) — competence- and benevolence-based trust mediate tie-strength → useful knowledge transfer; once trust is controlled for, weak ties re-emerge as valuable, [[ontario-trades-info-ecosystem-five-layer]], [[brokers-in-ontario-trades-three-roles]].
2. Referral mechanics — [[schmitt-skiera-van-den-bulte-2011-16pct-ltv-premium]], Van den Bulte, Bayer, Skiera, Schmitt 2018 (JMR) — referred-customer LTV decomposes into "better matching" (temporary) + "social enrichment" (persistent), [[trusov-bucklin-pauwels-2009-wom-elasticity-053]], [[warm-intro-hierarchy-eight-levels-ontario-trades]], Expectation transfer + retention by tie strength — strong-tie referrals carry expectation specificity + active social-enrichment retention; weak-tie referrals churn at near-baseline rates.
3. In-group vs market reputation — [[rindova-2005-being-good-vs-being-known]], [[in-group-vs-market-reputation-trades-distinction]], [[opacity-specificity-tradeoff-reputation-channels]], [[why-in-group-dominates-three-compounding-mechanisms]], [[dulleck-kerschbamer-sutter-2011-liability-verifiability-credence-goods]].
4. Accreditation and peer-network infrastructure — [[red-seal-endorsement-individual-journeyperson-credential]], [[bta-contractor-growth-method-600-contractors]], EOS Worldwide — "Over 200,000 businesses around the world"; 800-1000+ implementers; Traction (Wickman 2007/2011); heavy Ontario contractor adoption >$3M revenue, Vistage (45K members in 35 countries) + The Alternative Board (franchised local-board model) — generalist CEO peer-advisory, not trades-specific; both serve Ontario GCs, [[peer-coached-contractor-persona-synthesis]], [[events-as-trust-witnessing-not-lead-generation-infrastructure]]; existing entries referenced: [[hcra-launched-feb-1-2021-replaces-tarion-regulator]], [[hcra-ontario-builder-directory-is-load-bearing-not-hba]], [[tarion-established-1976-onhwpa-not-for-profit]], [[gold-seal-cca-1991-origin-and-purpose]], [[renomark-origin-bild-2001-transfer-chba-2024]], [[bild-gta-2008-merger-1000-members-2023-ohba-tension]], [[breakthrough-academy-coaching-self-reported-outcomes]].
5. Earned vs purchased reputation — [[spence-1973-job-market-signaling]], [[signalling-theory-applied-ontario-credential-stack]], [[best-of-houzz-low-cost-engagement-driven-signal]], [[guildmaster-guildquality-audit-verified-survey-signal]], [[audit-gap-problem-pooling-equilibrium-degradation]], [[sophisticated-gc-five-question-credential-heuristic]].
6. Research gaps — [[trust-networks-brief-research-gaps-may-2026]].

Relation to Briefs #1 and #2

• Brief #1 ([[research-brief-psychology-gc-marketing-aversion-may-2026]]): This brief refines the operational interpretation. If Brief #1 is read as "build better marketing assets to overcome aversion," that read is wrong. The aversion is downstream of the credence-good problem ([[dulleck-kerschbamer-sutter-2011-liability-verifiability-credence-goods]], [[marketing-services-as-credence-good-for-gc]]) — which is not solved by asset quality but by liability + verifiability + in-group reputation. Brief #1's empirical claim (48.7% of builders source >50% of sales from referrals — [[apb-sorci-2024-48-7pct-referrals-half-of-sales]]) points the same direction.
• Brief #2 (Research brief: risk aversion, loss aversion, and post-failure decision patterns in GC and trades-business decision-makers (May 2026)): Amplifies, not contradicts. Brief #2's "social proof works only when the reference group is recognizably proximate" ([[goldstein-cialdini-griskevicius-2008-provincial-norms-hotel-towel]], [[rule-radius-matched-peer-social-proof-only]]) is operationalized here as: proximity in Levin-Cross terms is trust, not closeness. Proximity-by-shared-cohort (BTA, EOS, RenoMark, BILD) is stronger than proximity-by-geography. A Toronto GC referred by a BTA cohort peer in Calgary is a closer tie than a Toronto GC referred by a geographically-adjacent acquaintance.