Marketing services are a near-paradigmatic credence good for a GC buyer

Claim: Marketing services to a GC satisfy all three Darby–Karni / Dulleck–Kerschbamer criteria for a credence good ([[darby-karni-1973-credence-goods-framework]]):

1. The GC cannot evaluate the technical quality of the marketing work the way he can evaluate framing or rough-in plumbing.
2. The GC cannot easily verify whether reported outcomes (leads, traffic, brand lift) are attributable to the agency's work or to background market factors.
3. The GC cannot in many cases verify after the fact whether the work itself was actually performed at the level represented.

Combined with GC risk aversion ([[kahneman-tversky-prospect-theory-loss-aversion-2to1]]), the credence-good structure produces a market in which the default rational stance is skepticism, not engagement. Reputational and referral mechanisms partially compensate — which is precisely why peer-referred marketing vendors close at far higher rates than cold-pitched ones, and why [[apb-sorci-2024-48-7pct-referrals-half-of-sales]] dominates lead-sourcing data.

Confidence: Verified (application of the academic framework to the domain).

For Candid — the strategic move: Concede the credence-good problem instead of denying it. Marketing vendors who say "trust us, the work is real" are arguing against information the buyer already has ([[ftc-homeadvisor-angi-jan-2023-7-2m-order]]). Candid says the opposite: "marketing is hard to inspect; here is exactly how we make it inspectable." Published process, published reporting, named-deliverable contracts, refusal-to-quote-cold — all forms of conceding the problem rather than pretending it doesn't exist. Operationalized as [[rule-concede-credence-good-problem-make-marketing-inspectable]].