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.
Landing page treated as a targeted revision in one record and as a possible layout rebuild or reuse of the 20-minute page in another
On the record
The existing landing page is being revised against specific client comments rather than fully redesigned
Arrived since
Options under consideration for the landing page are a layout/design change or reusing the 20-minute page structure changed to say 60 minutes
Why it matters: One record commits to minimal revision, the other opens a structural change. Since landing page changes are deliberately deferred to avoid confounding creative results, acting on the wrong record could invalidate the creative test window that is also on record.
Drafted options — clicking fills the note, it does not submit
Sending domain given as mail.controlcareer.com while every other domain record uses controlaltcareer.com
On the record
Sending emails from a subdomain (e.g. mail.controlaltcareer.com) rather than the root domain is the recommended setup; all calendar links now run on the controlaltcareer.com domain
Arrived since
Email sending is configured in the CRM with SPF and DMARC records in place, with a sender name of Jennifer and an underlying sending domain of mail.controlcareer.com
Why it matters: If the sending domain genuinely differs from the brand domain by a missing 'alt', SPF and DMARC alignment and domain reputation work are being applied to the wrong host, and deliverability remediation would silently fail. If it is a transcription error, propagating it into the knowledge base will mislead whoever configures DNS next.
Drafted options — clicking fills the note, it does not submit
Graphic creatives turned off campaign-wide, where the existing record pauses graphics only as part of the motherhood ad set
On the record
The motherhood-targeted ad set and its associated graphics were paused for producing consistently low-quality leads
Arrived since
Graphic (non-video) ad creatives have been turned off and will not be used at the start of the new campaign
Why it matters: The existing pause is qualified to one ad set and one reason. The incoming drops that qualifier and applies to all graphic creatives. If merged, the knowledge base would record that graphics were abandoned for producing low-quality leads, which is a broader and unverified conclusion about a whole format.
Drafted options — clicking fills the note, it does not submit
Forward-looking claim that manual audience targeting will not be available in 2027
On the record
Agency strategy proposal is to focus on AI-based targeting and creative diversity in alignment with Meta's Andromeda update
Arrived since
Agency position is that manual audience targeting is no longer significant and will not be available as an option going into 2027
Why it matters: The first half restates an approved position; the second half is a prediction about Meta's product roadmap presented as fact. If it reaches a client deck or an ad, the brand is making an unverifiable platform claim it cannot substantiate, and the strategy rationale collapses if Meta retains manual targeting.
Drafted options — clicking fills the note, it does not submit
Three irreconcilable counts of website lead volume reported for overlapping periods
On the record
Recorded 27 total inbound website leads (all sources combined) in the period from approximately 16 August to 15 September 2026
Arrived since
Meta reported six 'applications submitted' website leads while internal data showed three or four; Meta Event Manager showed roughly 10 website lead events against roughly 10–11 Squarespace form submissions, of which three carried Meta UTM attribution
Why it matters: Meta is over-reporting website leads relative to internal data by roughly a factor of two, and the newly configured CAPI, pixel removal, UTM hidden fields and 'contains' conversion rule all landed around the same time. Until the source of the gap is identified, cost-per-lead figures are unreliable and the campaign may be optimising toward phantom conversions. The 27-lead figure is a different measure and period and should not be read as contradicting these.
Drafted options — clicking fills the note, it does not submit
Enquiry-source list restated with 'Canvas' dropped and new channels added
On the record
Acquires enquiries through multiple non-Meta channels, including Google search and a third-party lead platform transcribed as 'Canvas', in addition to paid Meta ads
Arrived since
Receives enquiries from non-Meta sources including Google Ads, organic and walk-ins, in addition to paid Meta ads
Why it matters: 'Google search' and 'Google Ads' are not the same channel — one is organic, one is paid — and the incoming statement separates 'organic' out while naming Google Ads, which conflicts with the same batch reporting that Google Ads has never launched. The third-party platform 'Canvas' has also disappeared from the list without explanation. Treated as drift for the purposes of updating the record, but the channel boundaries moved enough to warrant a check before any attribution work relies on this list.
Drafted options — clicking fills the note, it does not submit
The second centre's location is recorded under four mutually exclusive names in a single claim
Arrived since
Second centre location recorded variously as Bishan, Bukit Manis, Bitima and 'Bit Marsh'
Why it matters: These cannot all be true at once and no passage of time explains the difference, so this is a contradiction rather than drift. Bishan and the Bukit Timah area are different parts of Singapore; publishing the wrong one sends prospective parents to a centre that does not exist and makes every location-targeted ad unverifiable. 'Bukit Manis', 'Bitima' and 'Bit Marsh' read as transcription variants of one name, but the system must not auto-resolve a location on phonetic resemblance — that is exactly the surface-similarity merge this step exists to prevent.
Drafted options — clicking fills the note, it does not submit
Email list of 9,000 is referred to as a WhatsApp-reachable lead list
On the record
Existing customer/contact database of 9,000 email addresses
Arrived since
WhatsApp reactivation of a 9,000-lead list
Why it matters: An email address is not a WhatsApp-reachable phone number. Treating the two as the same list silently assumes phone coverage across 9,000 contacts, which drives both the revenue projection and the sending ramp plan. If coverage is lower, the whole plan is built on a number that was never verified.
Drafted options — clicking fills the note, it does not submit
Conflicting account of whether the client is consulted before changes are made
On the record
Agency's media buying working style is to inform the client of the plan and proceed unless the client objects
Arrived since
Agency's working style is to make landing page changes first and inform the client afterwards, and to run such changes as split tests
Why it matters: One version gives the client a chance to object before a change ships; the other does not. Getting this wrong on a live landing page means unannounced changes to a page the client believes is fixed, and a trust problem that is expensive to repair.
Drafted options — clicking fills the note, it does not submit
Projected $10,000 single-day revenue from WhatsApp list reactivation is asserted without evidence
On the record
Existing customer/contact database of 9,000 email addresses
Arrived since
Agency states WhatsApp reactivation of a 9,000-lead list could unlock around $10,000 in one day after roughly two weeks of sending
Why it matters: This is a forward-looking revenue figure attached to a real list size. If it enters the knowledge base as a fact rather than a projection, it can surface in client-facing materials as a promise the agency has not demonstrated for this list.
Drafted options — clicking fills the note, it does not submit
Agency states two different core specialisations
On the record
Agency identifies its specialty as back-end work
Arrived since
Agency identifies its specialisation as paid ads, defined by acting immediately when it sees a winning signal
Why it matters: The speciality claim is used in positioning and proposals. If the knowledge base holds two different answers, outbound materials will describe the agency inconsistently and the client will not know what they are buying.
Drafted options — clicking fills the note, it does not submit
Incoming statement says only confirmed resale leads are forwarded, while the existing record says raw leads go directly to the client
On the record
Leads are vetted manually to verify genuine requests for floor plans and pricing
Arrived since
Automated lead filtering checks user identity and interest in packages before a lead is forwarded
Why it matters: Whether vetting is manual or automated determines how much team time lead handling consumes and whether the stated filtering criteria can be relied upon. Both are recorded as current, with different criteria.
Drafted options — clicking fills the note, it does not submit