Pricing checks that people do not learn to ignore

The failure mode for any automated pricing check is not missing things. It is training the team to click past it. A tool that flags sixty trivial items per banner ends up ignored on the three that carry campaign risk. An offer pre-check has to earn its place on every run, which means being deliberate about what it blocks on, what it merely warns about, and what it refuses to judge at all. The design constraint is attention, not coverage — and with a one-line promo, the ratio matters even more than in a long email.

Block on what is certainly wrong

Some constructions are facts, not opinions. A percentage claim that contradicts the two prices printed beside it is arithmetic, not style: if the draft says "90% off" next to $120 → $29, the real discount is about 76%, and the deception standard at 16 CFR Part 251 measures claims by what they convey to a reasonable consumer — the conveyed number simply is wrong. A "was" price equal to or lower than the sale price is the same class of fact: the former-price guide §233.1 frames a reference price as a representation about actual prior offering, and a reference price at or below the current price represents something the promotion itself disproves. These checks never produce false positives about whether the math disagrees; the only honest uncertainty is how a given reviewer will weigh the pattern. Block here, before publish, because these are the findings that escalate into challenge-level consequences.

Warn on what is probably wrong

Patterns the guides agree on, but that sensible marketers sometimes break on purpose, belong in the middle tier. "List price" or "compare at" phrasing is not automatically deceptive — the §233.2 guidance on comparisons turns on whether a bona fide market price exists at that figure, which the machine cannot see. A "free" offer without an accompanying cost-or-condition line is the classic warning: §233.3's concern is the price being increased to absorb the "free" item, and often the honest fix is one disclosure sentence, not a killed campaign. ALL-CAPS discount bursts, double exclamation stacks, "hurry" vocabulary, and countdown phrasing behave the same way — challenge-raising signals in promo traffic, not crimes. Scarcity words deserve a warning slot for a rarer reason: "only 3 left" may be literally true and unprovable from the paste. Warn, quote the sentence, let a human decide — context the machine cannot see (inventory tables, pricing history) often justifies what the pattern alone condemns.

Leave alone what cannot be decided mechanically

Whether your "was $120" was actually offered openly for a reasonable period — that lives in your order history, outside the pasted text, and an honest checker refuses to pretend otherwise. OfferPrecheck never connects to a store, an ad account, or a price-history service; it lists substantiation as a labeled advisory and stops there. Whether the vendor's "list price" reflects a bona fide market — invisible from a draft. Whether a regional law adds requirements beyond the US guides — EU, UK and Australian consumer-law positions appear only as context advisories, never as rulings, because the primary texts outside the US were not reachable from the build environment that produced the rule set (the exclusion log lives with the sources file). A tool that pretends otherwise produces confident noise, which is worse than silence: it burns the credibility you need for the checks that genuinely know things. This is also why an honest pre-check refuses to output a "compliance probability." Nobody can compute it — enforcement is discretionary — and pretending otherwise is the check becoming the deceptive pattern it flags.

A workload that stays sane

Teams that keep a pricing check alive run it on a fixed, boring cadence: one scan per draft, before the first publish, never after the campaign calendar is committed. The scan is fast precisely because it is mechanical — parse the price pair, run the arithmetic, match the published-factor vocabulary — and it takes seconds and produces the same ranks every time. What the cadence protects is the marketer's attention: if the block tier is already cleared by the time a human reads the draft, the human spends their effort on the judgment calls the tool deliberately refused to make. That division — machine on arithmetic and patterns, human on evidence and intent — is the only arrangement that has held up in practice. Teams that reverse it (humans recomputing discount percentages at midnight, tools pronouncing on pricing history) produce exhausted marketers and untrustworthy scores.

What to measure instead of compliance theatre

A single pass/fail verdict is the enemy: it hides the one finding that matters behind fifty that don't, and it invites bargaining ("we ran that number last season"). Named, tiered findings survive disagreement — someone can argue with a rank, but nobody can argue with a quoted sentence, a parsed price pair, and a named factor. If you must report a number, report the count of open HIGH findings, because that is the only figure with an obvious action attached: zero, or the draft does not publish. Dashboards and "offer health scores" sound like governance and function as decoration; the fix checklist, closed line by line, is the actual control.

Tiering is the product

What teams actually want from a pricing check is not more findings; it is a trustworthy ordering of them. Severity tiers — arithmetic contradictions and reference-price patterns first, disclosure and substantiation warnings second, tone noise last — are what keep the check on the desk instead of on the mute list. OfferPrecheck separates these tiers rather than collapsing them into one verdict, because a single undifferentiated signal is what gets muted. And every tier stays reviewable: quoted evidence, named factor, suggested edit — nothing silently rewritten, nothing published on your behalf.

Decision-support only: findings quote your copy and the published pricing guides. You make the call before anything goes live. Heuristic pre-check — not a pricing-law certification or legal advice; no historical-price verification, store or ad-account connection, or platform-policy measurement performed.