House of Spoils — Performance Decline Analysis

01 — Headline

Peak Revenue / Day

$25.0k

Sep 14–18 avg (5 days)

Verified: $125,161 ÷ 5

Current Revenue / Day

$14.4k

↓ Oct 1–6 avg (6 days)

Verified: $86,420 ÷ 6

Revenue Decline

−42%

↓ $25,032 → $14,403 / day

Shopify paid orders, net of tax

High-Value Orders / Day

−53%

↓ 5.8 → 2.7 orders ≥ $1,500

These drove 60–70% of peak-day revenue

Real Demand Signal

+55%

↑ Rev/session $4.97 → $7.68

Demand is NOT the problem

Qualified Traffic Lost

−27%

↓ 48,261 → 35,426 sessions

Ad delivery degradation, not demand

Bottom line: Revenue per qualified visitor is up 55% — real buyers are still converting well. The problem is that ad delivery is sending fewer real visitors and more junk clicks than 4 weeks ago. This is entirely an ad-ops issue, not a product or demand issue.

02 — Revenue Trend

Daily Net Revenue — Sep 9 to Oct 6, 2026

Shopify paid orders, net of tax, LA timezone. 28 days.

Sep 14–18 peak window Sep 9–22 (Period A) Sep 23–Oct 6 (Period B)

03 — High-Value Orders

Orders ≥ $1,500 Net — The Missing Revenue Driver

On Sep 14–18, orders ≥ $1,500 accounted for est. 60–70% of daily revenue. These collapsed from 5.8/day to 2.7/day — driving most of the revenue gap. ✓ Verified

Orders ≥ $1,500 net (LA timezone). On strong Sep days, each of these orders contributed $1,500–$5,000+ to daily revenue.

04 — Root Cause

Two Compounding Ad-Delivery Failures

Each mechanism is independently evidenced. Together they explain the revenue gap. A third issue — Dubai junk clicks — pollutes tracking data but has minimal revenue impact (footnote below).

Mechanism 1

IG Feed Audience Fatigue

HOS's highest-quality traffic source — Instagram Feed — started costing materially more per click around Sep 16, while delivering fewer clicks per dollar. This is the primary long-term traffic quality degradation.

✓ Verified — Meta placement breakdown

CPC Sep 9–15 $1.32

CPC Sep 30–Oct 6 $2.06 (+56%)

Clicks/day Sep 9–15 2,053/day

Clicks/day Sep 30–Oct 6 1,139/day (−45%)

Implication 36% fewer IG clicks per $

Mechanism 2

Main US Campaign Paused Oct 1

Clay paused the account's largest campaign ($3,500/day US TOF CBO). Replacement Q4 campaigns ran at $2,844/day — a $650/day shortfall — and cost-cap constraints meant slow ramp.

✓ Verified — Meta campaign insights

Old campaign budget/day ~$3,500

Old campaign — Oct 1 spend $126 (effectively zero)

Q426 campaigns Oct 1–6 $17,062 total

Expected from old campaign ~$21,000 (6 days)

Net spend gap Oct 1–6 −$3,938

Footnote — Data Hygiene

Dubai Junk Clicks: Tracking Noise, Not Revenue

Dubai has always been a minimal revenue contributor — this is NOT a driver of the decline. But its ~9× scale-up (Sep: $591 → Oct: $5,190 over 14 days) with Advantage+ routing to $0.10–0.12 CPC FB Reels/Overlay inventory floods analytics with accidental-tap sessions, making traffic and bounce data unreliable. Worth fixing for data hygiene only.

⚠ Strong Inference — CPC pattern + Aug precedent; bounce not directly verified

FB Reels CPC (Dubai) $0.11

FB Reels Overlay CPC $0.10

Total junk-placement clicks 18,888 (Sep 23–Oct 6)

Dubai ROAS Period A → B 1.47× → 0.75×

Oct 6 site-wide bounce 97–100% across all channels

Click-to-Session Quality Collapse

Meta click count held near flat (−3.6%). GA4 paid social sessions dropped −26.6%. The gap is junk clicks that don't become real site sessions. ✓ Verified

Period	Meta Clicks	GA4 Sessions	Quality Ratio	Revenue / Session
Sep 9–22 (Period A)	81,324	48,261	59.3%	$4.97
Sep 23–Oct 6 (Period B)	78,412	35,426	45.1%	$7.68
Delta	−3.6%	−26.6%	−14.2 pts	+55% ↑ demand

05 — What Was Ruled Out

Not a Demand Problem

Four alternative explanations investigated and eliminated.

✓ Site Breakage

PDPs, cart, checkout, and theme verified working throughout the period. No downtime incidents.

✓ Ad Spend Cut

Daily spend held within 7% across all 4 weeks. Revenue didn't track a spend drop.

✓ Demand Collapse

Revenue per qualified session rose every week: $4.97 → $5.99 → $6.26 → $7.68. Real buyers convert better, not worse.

✓ Seasonality

Approaching Q4 — HOS should see neutral to tailwind. Revenue-per-session trend is inconsistent with demand headwinds.

✓ Price / Product Change

AOV held flat ($977 Period A vs $971 Period B). No known pricing changes. Strong AOV on strong days in both periods.

✓ Attribution / Lag

Revenue decline is real Shopify orders, net of tax. Not an attribution artifact. Shopify data is 100% accurate at time of pull.

06 — Actions for Clay

Recommendations

Ordered by expected revenue impact. First two have zero downside risk.

1

Loosen or Raise Cost Caps on Q426 Campaigns

The three Q426 replacement campaigns spent $17,062 Oct 1–6 vs the ~$21,000 the paused campaign would have spent — a $3,938 gap. Cost-cap constraints are limiting ramp speed. The Evergreen Cost Cap campaign especially (ROAS 0.26× in ramp) appears to be cap-constrained, not underperforming by intent. Review cap levels; consider loosening to ~$1.5× target CPA to allow the learning phase to close faster.
Urgent
2

Fresh Creative for IG Feed

IG Feed CPC rose from $1.32 to $2.06 (+56%) over 4 weeks — a textbook audience fatigue signal. The active creative set is exhausted. Introducing fresh static imagery or video creatives on IG Feed is the fix. Without new creative, the CPC trajectory will continue rising, further compressing qualified traffic.
High Priority
3

Audit UK TOF Campaign

UK TOF ROAS collapsed from 1.24× (Period A) to 0.00× (Period B) on $3,167 spend — zero attributed revenue. Either attribution failure or the campaign exited its learning phase into poor delivery. Warrants a direct check.
Investigate
4

Monitor Oct 6 Junk Wave

Oct 6 showed near-universal 97–100% bounce across all GA4 channels (7,129 total sessions). 4,934 Unassigned sessions at 100% bounce; Cross-network 3,813 at 99.3% bounce. If this recurs, consider Cloudflare bot filtering on the same datacenter ASNs as the Aug 31–Sep 1 Singapore bot wave.
Monitor
5

Data Hygiene: Exclude FB Reels & Reels Overlay from Dubai Campaign

Not a revenue driver — Dubai revenue has always been minimal. But Advantage+ routing $0.10–0.12/click to FB Reels/Overlay generated 18,888 accidental-tap clicks in 14 days, polluting analytics (same pattern as the Aug IG Stories carousel fix). Switch Dubai to manual placements (IG Feed + IG Reels + IG Stories) to keep tracking data clean. Zero revenue risk.
Data Hygiene

07 — Appendix

Full Data Tables

Shopify Daily Revenue & Orders Sep 9–Oct 6 · Paid orders · Net of tax · LA timezone

Date	Orders	Net Revenue	AOV	≥$1,500 Orders
Sep 9	17	$16,144	$950	3
Sep 10	14	$9,785	$699	1
Sep 11	24	$19,363	$807	5
Sep 12	20	$14,068	$703	2
Sep 13	19	$15,209	$800	3
Sep 14	20	$25,333	$1,267	5
Sep 15	18	$23,697	$1,317	4
Sep 16	19	$23,029	$1,212	6
Sep 17	21	$28,050	$1,336	7
Sep 18	24	$25,052	$1,044	7
Sep 19	10	$6,794	$679	1
Sep 20	19	$20,176	$1,062	5
Sep 21	22	$19,196	$873	4
Sep 22	23	$17,871	$777	2
Sep 23	19	$16,623	$875	3
Sep 24	12	$11,923	$994	3
Sep 25	15	$7,792	$520	0
Sep 26	17	$10,752	$632	1
Sep 27	29	$19,018	$656	3
Sep 28	19	$26,315	$1,385	5
Sep 29	20	$31,324	$1,566	5
Sep 30	19	$33,712	$1,774	7
Oct 1 ← US TOF paused	20	$14,896	$745	3
Oct 2	12	$15,793	$1,316	4
Oct 3	24	$17,462	$728	2
Oct 4	16	$12,842	$803	3
Oct 5	21	$12,058	$574	1
Oct 6	8	$13,370	$1,671	3

Meta Account-Level Daily NY timezone · act_16731102

Date	Spend	Clicks	CPM	CPC
Sep 9	$5,587	6,119	$20.78	$0.91
Sep 10	$5,092	4,902	$22.74	$1.04
Sep 11	$5,321	5,581	$22.66	$0.95
Sep 12	$5,767	5,949	$25.11	$0.97
Sep 13	$7,369	8,138	$24.10	$0.91
Sep 14	$4,350	5,139	$22.16	$0.85
Sep 15	$5,547	5,813	$22.70	$0.95
Sep 16	$5,348	6,850	$21.17	$0.78
Sep 17	$4,953	5,486	$24.75	$0.90
Sep 18	$5,098	5,689	$24.99	$0.90
Sep 19	$4,848	4,738	$15.45	$1.02
Sep 20	$5,739	6,087	$25.58	$0.94
Sep 21	$5,190	5,308	$22.76	$0.98
Sep 22	$5,287	5,525	$21.19	$0.96
Sep 23	$5,501	6,582	$12.98	$0.84
Sep 24	$4,973	4,667	$19.47	$1.07
Sep 25	$5,914	5,139	$22.70	$1.15
Sep 26	$4,521	4,033	$22.69	$1.12
Sep 27	$6,130	5,421	$26.13	$1.13
Sep 28	$4,635	4,964	$21.28	$0.93
Sep 29	$4,423	5,232	$15.26	$0.85
Sep 30	$5,294	7,258	$15.12	$0.73
Oct 1 ← TOF paused	$5,262	4,874	$15.25	$1.08
Oct 2	$3,664	4,548	$11.30	$0.81
Oct 3	$4,205	4,123	$17.92	$1.02
Oct 4	$4,919	5,208	$14.66	$0.94
Oct 5 ← Junk spike	$5,996	8,066	$13.91	$0.74
Oct 6 ← Junk spike	$4,603	8,299	$12.58	$0.56

Dubai Campaign Placement Breakdown Sep 23–Oct 6 · CE_Dubai_TOF_CBO_Q326

Placement	Spend	Clicks	CPC	Signal
instagram/feed	$1,521	1,211	$1.26	Quality
instagram/instagram_stories	$897	533	$1.68	Quality
instagram/instagram_reels	$724	1,177	$0.61	Acceptable
facebook/facebook_reels	$705	6,711	$0.11	Junk
facebook/feed	$589	4,853	$0.12	Junk
facebook/reels_overlay	$548	5,256	$0.10	Junk
facebook/instream_video	$162	2,068	$0.08	Junk
Total junk placements	$2,004	18,888	$0.11 avg

⚠ Inference "Junk" classification based on CPC pattern ($0.08–0.12 = accidental-tap inventory) and Aug IG Stories carousel precedent. Bounce rate for these placements not directly verified in GA4 (would require Dubai UTM/source matching). Zero attributed purchases from FB Reels or Reels Overlay.

Methodology & Timezone Notes Data definitions · Sources · Caveats

Shopify revenue = total_price − total_tax for financial_status=paid orders only. Excludes refunded/partially_refunded orders. LA timezone (America/Los_Angeles, UTC−7 in Sep/Oct 2026). Paginated 3 pages × 250 limit = 521 total orders verified.

Meta data uses NY timezone (America/New_York, UTC−4 in Sep/Oct 2026). Date alignment: NY dates in Sep/Oct 2026 match LA dates (NY is 3h ahead but both standard/daylight). Attribution window: 7-day click / 1-view. Meta-attributed ROAS is directional only — treated as signaling, not truth. MER = Shopify net revenue ÷ Meta spend (same calendar day).

GA4 sessions use sessionDefaultChannelGroup. Property 376958515. "Paid Social" channel captures Meta traffic. Singapore bot wave (Sep 21–28) inflated "Direct" channel — not included in paid social analysis. Session counts are GA4's own session definitions (not screenPageViews).

Evidence standards: "✓ Verified" = claim was re-queried from source API and matched within 2%. "⚠ Strong Inference" = claim is logically supported by direct evidence (CPC pattern, click trajectory, August precedent) but the specific metric (bounce rate by placement) was not directly verified against GA4 or landing page data.

Data pull timestamp: All figures queried Oct 7, 2026, approximately 08:10 UTC+2. Meta data for Oct 5–6 may shift slightly as attribution finalizes (~48h lag). Shopify data is exact at time of pull.