📚 CANI Session Report
Midas — Meta Media Buying Agent | March 18, 2026 | Live Bearded
5
Stale Beliefs Identified
🔄 Stale Beliefs — Identified & Updated
Belief #1 CRITICAL HIGH CONFIDENCE it's stale
Old belief: iOS 14.5 is the primary attribution and tracking challenge to diagnose.
Why it's wrong: iOS 14.5 launched in April 2021. It's been 5 years. The ecosystem has adapted — CAPI is table stakes, not a differentiator. More recent disruptions (iOS 17 Link Tracking Protection, ATT normalization, Meta's own attribution model changes) have layered on top. Framing every attribution issue through an iOS 14.5 lens is outdated.
Updated belief: iOS 14.5 is the baseline we assumed. The active challenges in 2026 are iOS 17 LTP (which strips URL parameters including UTMs on Safari), Meta's move toward modeled/aggregated conversions, and the divergence between Meta-reported ROAS and actual Shopify revenue. CAPI is required but not sufficient — EMQ score and deduplication accuracy are the real diagnostics now.
Action: Stop citing iOS 14.5 as a current issue. Start diagnosing by EMQ score, CAPI dedup rate, and MER vs. Meta ROAS gap. Add iOS 17 LTP to the tracking health checklist.
Belief #2 CRITICAL MEDIUM CONFIDENCE it's stale
Old belief: Interest-based targeting is a legitimate default prospecting strategy.
Why it's wrong: Meta's algorithm has significantly matured. The practitioner consensus I'm seeing in my sources (Foxwell, Plofker, Holiday) points toward broad targeting (no interest restrictions) + strong creative as the primary prospecting strategy in 2026. Interest targeting may be restricting Meta's ability to find buyers and increasing CPMs unnecessarily. Advantage+ audiences have further shifted the paradigm.
Updated belief: For LB's account at current spend levels (~$6K/day), broad targeting with strong creative is likely more efficient than interest stacking. The creative IS the targeting. Interest targeting may be appropriate for initial testing phases or very niche angles, but shouldn't be the default structure.
Action: Flag this for the deep dive. Need to see LB's current ad set targeting setup. If interest-heavy — this is likely a structural contributor to the below-breakeven ROAS. NEEDS VERIFICATION FROM MINK
Belief #3 HIGH PRIORITY LOW CONFIDENCE — no current knowledge
Old belief: CBO (Campaign Budget Optimization) is generally preferred over ABO (Ad Set Budget Optimization) for most campaigns.
Why it might be wrong: Meta has evolved its budget allocation algorithms substantially. Advantage+ campaigns have introduced a third paradigm. The CBO vs. ABO debate has evolved — some practitioners now run ABO for testing phases and CBO for scaling, with Advantage+ for proven creatives. I have no current operational knowledge of what's working at LB's spend level.
Updated belief: Cannot hold a firm position on this without current account data and practitioner input. Suspending prior. Need to see LB's current campaign structure and what budget type is set on active campaigns.
Action: Add to deep dive checklist. Add CBO vs. ABO vs. Advantage+ to research queue. QUEUED FOR RESEARCH
Belief #4 HIGH PRIORITY LOW CONFIDENCE — unknown for LB
Old belief: 7-day click, 1-day view is standard attribution and presumably what LB is running.
Why it might be wrong: Attribution window defaults have changed multiple times since iOS 14. Meta may have auto-adjusted account settings. Some accounts are on 7-day click only, some 1-day click, some still 7-day click / 1-day view. The window in use dramatically affects reported ROAS. If LB is on a wider window, ROAS looks better than reality; narrower window, ROAS looks worse.
Updated belief: I don't know LB's current attribution window setting. This is a required data point before I can accurately interpret any ROAS number.
Action: Ask Mink: "What attribution window is set in Events Manager?" BLOCKING — need answer
Belief #5 MEDIUM PRIORITY LOW CONFIDENCE — zero operational knowledge
Old belief: Advantage+ Shopping Campaigns (ASC) are worth testing for eComm accounts.
Why it might be wrong: I know ASC exists and it's positioned as Meta's AI-first campaign type, but I have no operational data on how it performs for DTC grooming brands at LB's spend level. Some accounts have seen strong results; others have seen Meta shift budget in ways that hurt efficiency. I cannot make a qualified recommendation about ASC without actual knowledge.
Updated belief: Suspending judgment. Need to know: (1) Is LB currently running ASC? (2) What are practitioners reporting for DTC brands at $5K-10K/day spend? Add to CANI queue for proper research.
Action: Add to research queue as Priority 1. Don't recommend for or against ASC until researched. QUEUED FOR RESEARCH
✅ Best Practices — Updated This Session
| Area |
Old State |
New Rule Added |
| Attribution & Measurement |
iOS 14.5 framing |
Always check EMQ score + CAPI dedup rate first. MER (Total Revenue ÷ Total Ad Spend) is the true performance metric — Meta ROAS is a signal, not the verdict. iOS 17 LTP now strips UTM parameters on Safari — factor into click attribution gaps. |
| Targeting Philosophy |
Interest targeting assumed as default |
In 2026, creative IS the targeting. Default prospecting posture should lean broad (no interest restrictions) unless specific hypothesis warrants otherwise. Advantage+ audiences worth testing at LB's spend level. Flag any active interest-stacked campaigns for review. |
| Performance Benchmarking |
Platform ROAS as primary metric |
Always report MER alongside Meta ROAS. Meta ROAS alone is misleading — Klaviyo email sends, direct traffic, and attribution window selection all distort it. MER = Shopify revenue ÷ total Meta spend. Track weekly. Meta ROAS is a directional signal; MER is the truth. |
📡 Sources — Updated
✅ Added to sources.md (were missing, now added):
- Taylor Holiday — Common Thread Collective / @taylorholiday — MER-first framework, DTC Meta strategy, publishes real account-level benchmarks. Highly relevant for LB's current below-breakeven situation.
- Andrew Faris — AJF Growth / @andrewjfaris — DTC operator and advisor, current Meta mechanics, scaling frameworks for 7-8 figure brands. Essential voice for 2026.
Existing sources confirmed active and relevant: Jon Loomer, Andrew Foxwell, Nick Shackelford, Barry Hott, Cody Plofker, Zach Murray, Ben Heath
🔬 Research Queue — Updated
| Priority |
Topic |
Why It Matters for LB |
| P1 |
Advantage+ Shopping Campaigns — 2026 performance benchmarks for DTC |
May be relevant for LB's account structure. Cannot recommend without knowledge. |
| P1 |
Broad vs. Interest targeting — current best practice at $5K-10K/day spend |
Potential structural issue contributing to LB's below-breakeven ROAS. |
| P2 |
CBO vs. ABO vs. Advantage+ — 2026 framework |
Need informed position before making campaign structure recommendations. |
| P2 |
iOS 17 Link Tracking Protection — impact on Meta attribution |
May be contributing to ROAS measurement gap. Need to quantify. |
| P2 |
Creative fatigue benchmarks — DTC grooming brands, 2026 |
LB's creative refresh rate unknown. Need current frequency/CTR decline thresholds. |
| P3 |
Retargeting window optimization — 7 vs. 30 vs. 90 day performance |
LB has 398K+ customers. Retargeting window affects reach and frequency significantly. |
| P3 |
Competitor analysis — beard/grooming brands Meta creative strategy |
What are Beardbrand and others running? CPM benchmarking context. |
🎯 What I Need From Mink — In Order of Impact
1
Attribution window: What's set in Events Manager for the Purchase event? (7-day click / 1-day view? 7-day click only? 1-day click?) This affects every ROAS number I report.
2
CAPI status: Is server-side tracking running? What's the Event Match Quality score for Purchase in Events Manager?
3
Real MER: What is Shopify revenue ÷ Meta spend for the last 30 days? This is the number that tells us if Meta is actually profitable or not — not the 0.82x I'm seeing in Meta's dashboard.
4
Targeting structure: Are the active campaigns running broad (no interests) or interest-stacked? A screenshot of the ad set audience settings would answer this immediately.
5
When did ROAS start declining? Was it a gradual bleed over months or a sharp drop at a specific point? That answer changes the diagnosis completely.
🪞 Honest Assessment — Where My Knowledge Stands
| Area |
Status |
Confidence |
| Kill/Scale/Keep decision framework |
Solid. Logic is sound regardless of year. |
HIGH |
| ROAS diagnosis and root cause analysis |
Strong framework. Context-dependent execution. |
HIGH |
| Creative fatigue detection |
Good. Frequency + CTR signals are evergreen. |
HIGH |
| Current Meta campaign structure best practices |
Gaps acknowledged. Suspending priors on CBO/ABO/ASC. |
MEDIUM |
| Advantage+ Shopping Campaigns |
No operational knowledge. Will not recommend until researched. |
LOW |
| iOS 17 LTP attribution impact |
Aware it exists. Don't know LB's specific exposure. |
LOW |
| Current Meta audience targeting philosophy (2026) |
Directionally updated. Need to verify against LB's account setup. |
MEDIUM |
⚠️ Standing Alert — Not a CANI Issue
Yesterday's account ROAS: 0.82x. Spending $5,963, generating $4,909 in revenue. All three active campaigns are underwater. This is a live bleeding situation that exists independent of CANI gaps. The deep dive should happen now — CANI just means I'll bring better questions to it.