Review
1444 facts across 22 clients. Two questions each: can we stand behind it, and can an argument be built from it.
Nothing here blocks anything. Strategy already runs on these — approving a fact only tells it which to reach for first. Ten minutes on your biggest client is worth more than clearing the list.
Two things the system found that disagree, and it refused to pick. Each carries both claims and the options the critic drafted — settling one writes how it was settled, because the next person to hit the same contradiction reads only that.
$19,000 stated as both a fixed and a starting price
On the record
Resale Reno Package from $19,000
Arrived since
Resale Reno Package for just $19,000
Why it matters: Both appear in the same script. In a paid ad a fixed price and a starting price are materially different promises, and the riskier reading is the fixed one — a homeowner quoted more later has a legitimate complaint.
Drafted options — clicking fills the note, it does not submit
A second warranty claim widens covered scope to 'everything installed' while shortening the term to two years
On the record
Lifetime warranty on carpentry hardware (Excel)
Arrived since
Two year workmanship warranty on everything installed
Why it matters: Two warranty statements with different scopes and different terms will be compressed downstream into either 'lifetime warranty' or 'two year warranty on everything', both of which misstate coverage in paid advertising. Whichever survives, the qualifier that makes it true has to travel with it.
Drafted options — clicking fills the note, it does not submit
85% accuracy claim has no completed validation behind it
Arrived since
TCM tongue analysis accuracy stated at approximately 85%, with the blind research study still only planned
Why it matters: A numeric medical-adjacent accuracy figure is the single most likely fact to be lifted into a caption or ad as '85% accurate' with the 'approximately' and 'tongue analysis' qualifiers stripped. The only supporting evidence on record is an informal comparison against a clinical machine 'described as consistent', and the study designed to substantiate it has not run. Publishing this before validation is a health-claim exposure, not a marketing risk.
Drafted options — clicking fills the note, it does not submit
Partner's HSA and ISO certifications risk being read as the client's own
Arrived since
A prospective investor/partner operates a herbal medicine 'brewery' holding HSA and ISO certifications and a testing lab
Why it matters: Attribution is part of the fact. These are a prospective partner's credentials, and the partnership is not concluded. If this migrates into brand copy as 'HSA certified' or 'ISO certified facilities', the brand is asserting regulatory standing it does not hold — the highest-consequence version of broken telephone, and one that reads like a harmless rewording.
Drafted options — clicking fills the note, it does not submit
Three-month wear claim stated across the whole range while the range includes daily lenses
On the record
Product range consists of daily lenses and three-month lenses (incoming, same batch)
Arrived since
Lenses can be worn for up to three months
Why it matters: Read as written, this says every lens including the dailies can be worn for three months. Reusing it in an ad or product page would encourage single-use lenses to be worn for months, which is a health claim as much as a marketing one. The boundary moved while the sentence looks unchanged.
Drafted options — clicking fills the note, it does not submit
Knowledge base still asserts a visitable physical shop; the Bugis store has fully closed
On the record
Has a physical shop customers can visit
Arrived since
The Bugis physical store has fully closed / the brand has moved to an online and event-based model
Why it matters: The existing record is undated and present-tense, so anything generated from it can invite customers to a shop that no longer exists. This is not hypothetical: the website's appointment page is still routing bookings to Bugis Plus, and the only other premises — the warehouse — is confirmed not open to the public. Left unresolved, the brand has no walk-in location while its own site and knowledge base both say otherwise.
Drafted options — clicking fills the note, it does not submit
Team says it cannot perform eye checks, while the brand is publicly promising eye checks at its 1 October booth and owns a portable eye-test machine
On the record
The brand will have a physical booth at a K-pop event on 1 October where visitors can get an eye check and see the lens colours in person; The brand owns an eye-test machine; The eye-test machine is portable
Arrived since
The team cannot perform eye checks for customers who have no prior prescription or purchase record, and refers those customers to another optical shop before they can order online (and separately: 'The brand plans to offer eye checks at its event pop-ups')
Why it matters: An advertised eye check at a public event that the team cannot actually deliver is a promise made in market to first-time customers — exactly the people the referral rule turns away. Both versions are currently in the knowledge base and either could end up in event copy.
Drafted options — clicking fills the note, it does not submit
Knowledge base says customers can visit a physical shop; incoming says there is no physical retail outlet for the next six months
On the record
Has a physical shop customers can visit (also on file: 'The Bugis physical store has fully closed', 'The brand has moved from a physical Bugis store to an online and event-based model')
Arrived since
No physical retail outlet is planned for the next six months; the business is digital and event-based in that period
Why it matters: If the older capability record survives, marketing can invite customers to visit a store that does not exist — and the website's appointment page already directs them to Bugis Plus. That is a wasted trip and a credibility hit at the point of purchase.
Drafted options — clicking fills the note, it does not submit
Two current cost-per-lead figures for Haga that differ by roughly 10–20x
On the record
Haga cost per lead ranges from about $67 at its lowest to about $167 when conditions are poor
Arrived since
Haga is currently acquiring leads at $5 to $7 per lead
Why it matters: Both are stated as current for the same brand and the same measure, so no passage of time reconciles them. The figure that gets written down becomes the baseline every agency is briefed against: at $5–$7 the stated sub-$10 target is already met and the five-leads-per-$1,000 benchmark is trivial; at $67–$167 the same benchmark is demanding and the target is a long way off. The arithmetic points at a definitional split rather than an error — 20 leads/day is roughly 600/month, which against $2,000–$3,000 of ad spend gives about $4–$5 per lead, while the five-leads-per-$1,000 benchmark on total marketing spend implies a ceiling near $200 per lead. That is consistent with $5–$7 being ad-spend-only CPL and $67–$167 being fully-loaded CPL including labour and photography.
Drafted options — clicking fills the note, it does not submit
Product pregnancy-safety claim asserted with no substantiation source
Arrived since
Apondle's skincare products are safe for use during pregnancy
Why it matters: This is the first health/safety claim in the knowledge base and it arrives as an unsourced assertion. Skincare pregnancy-safety is a substantiation-required claim; once it sits in the record a scriptwriter can lift it into a paid video stripped of the incoming positioning guardrail, which is precisely the broken-telephone path to a false advertising exposure. Severity is set by consequence, not by how clear-cut the match is.
Drafted options — clicking fills the note, it does not submit
The 2K retainer is described both as part of a 7K total and as a monthly charge
On the record
Project financial structure of 7K total: 5K allocated to the MVP and a 2K retainer reserved for future feature development (same intake batch)
Arrived since
Payment schedule set to a 5k installment now, with the 2k monthly retainer deferred until a milestone check-in in September; The 2K monthly retainer is on hold until initial project milestones are complete
Why it matters: A one-off 2K and a recurring 2K/month cannot both be true of the same engagement, and no passage of time explains the difference — these were stated together. The recorded contract value is either 7K or open-ended, which affects invoicing at the September check-in and any figure quoted about the engagement.
Drafted options — clicking fills the note, it does not submit
The stated ~85% accuracy figure may rest on a number that is not an accuracy measure
On the record
TCM tongue analysis accuracy stated at approximately 85% (existing KB)
Arrived since
The confidence percentage currently shown in reports represents a ratio of potential constitutions rather than an absolute confidence score; the metric will be redefined to display statistical accuracy instead of the current proportion values
Why it matters: If the 85% originated from the on-report confidence figure, the brand is publishing an accuracy claim derived from a constitution-proportion ratio. That is a substantiation failure on a health-adjacent claim, and it is currently live in the knowledge base for downstream scripts and captions to reuse. Recency does not settle this either way — the figure needs provenance, not a newer restatement.
Drafted options — clicking fills the note, it does not submit