The CTV program is the only Frøya channel with causally measured lift: $206,493 of proven incremental revenue in a 33-day flight, 1.09x on an LTV basis. It has held 1.5–1.9x ROAS across a 3.3x range of weekly spend. The Converge warehouse now shows the store's new customers sliding while revenue holds at record levels — the current audience pools are saturating. The play: 3x the program to $600k/month by 11/19, pointed at the new audience pools where the lift is already proven to run — the 24 holdout states, the Southern Belt, and the pure-media markets.
A geographic holdout study, not dashboard attribution: the 26 states where CTV ran versus the 24 where it did not, against a clean 76-day pre-flight baseline. Two independent models, both positive.
The per-capita model measures $206,493 (0.88x direct). The revenue-share model — which makes no per-capita assumptions at all — measures $103,550 (0.44x), the conservative floor. The mechanical state-level read is $146,336. The disclosed band is $146k–$206k; every route is positive. On the store's own measured repeat behavior (30% repeat rate, 1.24x annual customer halo), the flight is 1.09x — profitable on the customers it wins.
The effect was largest right after launch (1.0474 in Aug 16–31) and was still above baseline in September (1.0307). It is not a launch spike. And it is not a September artifact: the Oct 12–18 week measured +2.97% (exposed states $1,397,078 vs. comparison $1,367,441 — a $40,332 difference) while the flight kept running.
The complete Frøya order history from the Converge warehouse — every order, every channel, Aug 1 to Nov 4, 2026: $49,022,792 in revenue, 488,512 orders, 269,803 new customers. The weekly series tells the story the dashboard numbers hide.
Revenue per day has held a $490k–$544k band all quarter — and the Oct 12–25 window is the revenue high of the record ($543,771/day). In that same window, new customers per day hit the low of the record (2,334), and new-customer share of orders fell to 44.3% from 59.9% pre-flight. The store is earning more while it acquires fewer new people — the existing customers and the repeat engine are carrying the number.
The window-by-window read, computed from the warehouse: pre-flight (8/1–15) 3,061 new customers/day → flight (8/16–9/17) 2,929 → post-flight (9/18–10/11) 2,805 → 10/12–25: 2,334 → 10/26–11/4: 2,720 (a partial recovery). The direction over the two months since the CTV flight reached its full 26-state footprint is down, and the mix is shifting toward repeat purchases.
The full timeline of the media program since the flight started, with the lift-report numbers each step produced — and the live CTV series that verifies them to the cent.
CTV flight launches — 26 states, $50k/week. 10.53M impressions delivered through Sep 17. Measured result: $206,493 incremental revenue, +3.70% per capita, 2,289 incremental orders — the lift study.
33-day flight window closes. The geo-lift study is published (froya-lift.thesingulariti.ai): 0.88x direct / 1.09x LTV, two models agreeing, lift still above baseline in the final 17 days ($79,027 in the Sep 1–17 slice alone).
Weekly lift report: CTV attributed $88,288.40 across 877 orders at 1.55x. The geo check on the same week: +2.97% (a $40,332 exposed-vs-comparison difference). Verified against the live CTV series: $88,288.40, 876 orders, 1.55x — to the cent.
Weekly lift report: $100,594.85 at 1.56x — attributed revenue up 13.8% week over week. Verified live: $100,594.84, 994 orders, 1.56x. The week's cost ran $64,642.50 — 29% above the $50k cap, which is where the room for the 3x ramp physically is.
Linear TV proposal delivered: $52,500 New York pilot ($50,000 media + 5% fee), one month, geographic-lift design. Decision deadline November 6 — the new-audience lever in the 3x plan.
The channel scales. ROAS held in the 1.51–1.92x band across a 3.3x range of weekly spend ($21,262 to $70,798). The high-spend weeks ran 1.53–1.62x — still above the 1.09x LTV break-even line and within the band the flight measured. CTV attributed revenue grew from $34,839/week (8/17) to $108,020/week (10/26) — 3.1x — while ROAS stayed in the band.
The honest attribution read. CTV-attributed orders are 3.8% of the store's new customers (8,543 of 223,882, 8/16–11/4), and the daily correlation between CTV spend and store revenue is r = +0.16. CTV is one channel inside a $500k/day store — it does not move the store's daily line, and it never claimed to. The causal proof is the geographic holdout in Section 01, not the daily correlation. What the daily series does show is that the store's new customers slid while the CTV footprint saturated — which is exactly the gap the 3x plan fills with new pools.
Four independent signals point the same way: the audience the current program reaches is being worked to its ceiling.
59.9% of orders pre-flight → 44.3% in Oct 12–25 (Converge warehouse, computed). The mix is shifting to repeat purchases — the acquisition engine is stalling while the revenue engine runs on existing customers.
1.63x (September) → 1.54x (October to date) — the current spend is sitting exactly at the 1.54x contribution break-even line (64.8% margin). Healthy, but the room in the current footprint is closing.
Median time-to-purchase by creative has increased over recent weeks (noted at the Nov 5 sync). The signature of a saturating audience: the same people, taking longer to buy.
Meta CPA is running at ~2.5x AOV — a P1 issue since September. New customers from Meta cost ~$142 each at first touch (96-day Converge measure) against a $100 AOV. The new-customer money belongs in the channel with proven incrementality.
The channel is proven (Section 01). The pools are the opportunity. Each one is documented in the lift study's own state-level data.
170.4M people who never saw a single impression — the control group is the unmined pool. Same proven mechanism, a population with zero prior frequency. This is where 3x reach comes from: the flight's 169.0M exposed population is the 1x; the holdout is the next 1x of geography.
1.19x above-baseline growth per impression — the best efficiency in the study, measured. Texas alone is the largest single state delta anywhere in the flight (+$117,612). These four states passed three tests: natural demand, the largest launch spikes (+12% to +27% in 16 days), and persistence into the freshest window.
Flat or declining before the flight, they spiked +12.4% to +18.5% anyway — CTV created demand there rather than riding it. The study's own words: "those are the scale-up markets." Demand the media creates is demand that does not saturate the way retargeting does.
$52,500 pilot ($50,000 media + 5% fee), one month, geographic-lift design. Reaches the women-40+ broadcast and cable environment the CTV buy does not — a new medium, not more frequency. Decision deadline November 6. It is the new-audience lever, not the fastest one.
The per-impression lift rate is measured over the 33-day flight. The pools change where the spend goes, not the mechanism. The Southern Belt's 1.19x per-impression efficiency is a measured number from the study, not a forecast. The holdout states and the pure-media markets are the study's own named expansion candidates. The only stated assumptions in the projection are the saturation discount (FLOOR case) and the no-saturation assumption for new populations (UPSIDE case) — both labeled, both bracketing the measured middle.
The current program spends ~$230k/month (measured: $249,190 in Aug, $227,591 in Sep, $237,230 in Oct) and produces $6,257/day of measured incremental revenue — 0.88x direct, 1.09x LTV. The 3x program is the same engine at $150k/week, ramped over three weeks to hit Black Friday at full rate.
| Scenario | Incremental rev / wk | New customers / wk | Direct iROAS | LTV iROAS | Basis |
|---|---|---|---|---|---|
| FLOOR — 3x spend, same 26 states, 25% saturation discount (stated) | $98,554 | ~765 | 0.66x | 0.81x | The flight's measured 0.88x rate, discounted for 3x frequency in saturated states |
| BASE — 3x spend in the proven high-response geographies | $157,500 | ~1,222 | 1.05x | 1.30x | Southern Belt + pure-media states at the measured 1.19x per-impression efficiency |
| UPSIDE — full 3x reach incl. the 24 holdout states | $163,500 | ~1,269 | 1.09x | 1.35x | New populations at the measured LTV-adjusted rate; no saturation in unexposed pools (stated) |
$681,975/month of incremental revenue — 3.6x the current program's measured $188,473/month — and ~5,292 new customers a month, 3.6x the current ~1,460/month measured run-rate. Even the FLOOR case — no new pools, a stated 25% saturation discount — is $428,670/month incremental at 0.66x direct, 0.81x LTV. The plan is not a bet on a new mechanism; it is the same measured mechanism, pointed where the lift already runs hottest, at three times the rate.
The new-customer fix is in the geometry, not the copy. The store's new-customer count fell 24% from its pre-flight level while revenue hit records — because the pools saturated. The 3x program adds 170.4M unexposed people (holdout states) and the demand-creating pure-media markets to the proven 26-state footprint. New customers follow new reach: 340/week measured at 1x → 765–1,269/week at 3x, with the BASE case at ~1,222/week.
Black Friday week starts 11/23. The ramp takes three weeks. The lift shows within one — the flight's own first 16 days ran 1.0474. The calendar is the constraint, not the media.
$75k/week (wk 1) → $110k/week (wk 2) → $150k/week (wk 3, from 11/19). Geographic allocation: Southern Belt + pure-media states first (measured 1.19x efficiency), then the 24 holdout states as delivery capacity opens. Full 3x four days before Black Friday week.
The $52,500 pilot is already proposed and costed ($50,000 media + 5% fee, one month, geographic-lift design). It is the new-medium lever — the women-40+ broadcast audience the CTV buy does not reach. The decision deadline is tomorrow; it runs in parallel with the CTV ramp, not instead of it.
With Frøya's finance team. It is the break-even line the financial case rests on (1.54x attributed ROAS = 1.00 at 64.8%). The measured 1.09x LTV read and the 3x projection both stand on it.
| Number | Source | Method |
|---|---|---|
| $206,493 lift · +3.70% · 2,289 orders · 0.88x / 1.09x | Frøya CTV Geo-Lift Study, froya-lift.thesingulariti.ai | Geographic holdout, 26 exposed vs. 24 control states, 76-day pre-flight baseline, Aug 16 – Sep 17. Two independent models (per-capita DID; revenue-share). Lift band disclosed: $146,336 mechanical / $206,493 per-capita. |
| Store revenue $49,022,792 · 488,512 orders · 269,803 new customers (96 days) | Converge warehouse, froya_converge.attribution_daily | Direct-session model, all channels, 2026-08-01 → 2026-11-04. Queried 2026-11-05; 96 days, daily grain verified. |
| New customers 3,061 → 2,334/day · share 59.9% → 44.3% | Converge warehouse (same table) | new_customer_order_events / placed_order_events, direct-session, by window: pre-flight 8/1–15, flight 8/16–9/17, 10/12–25, 10/26–11/4. |
| CTV attributed $88,288.40 (877 orders, 1.55x) · $100,594.85 (1.56x) · $64,642.50 week cost | Frøya weekly lift reports · verified live | Vision Bridge daily metrics (analytics.getDailyMetrics), pulled 2026-11-05: $88,288.40 / 876 orders / 1.55x and $100,594.84 / 994 / 1.56x — to the cent against the reports. |
| ROAS 1.51–1.92x across $21k–$71k/week · CTV rev $34,839 → $108,020/wk | Vision Bridge daily metrics (96 days) | Weekly aggregates, 12 full weeks. CTV orders = 3.8% of store new customers (8,543 / 223,882). |
| 1.19x Southern Belt per-impression · TX +$117,612 · OK/KS/IN/NY/NC pure-media | Geo-Lift Study, state-level tables | Measured state deltas vs. own Jun–Jul baselines; per-impression efficiency = delta / impressions. |
| $22.38 CPM · 10.53M impressions · $235,714 flight spend | Geo-Lift Study · Blockboard delivery file | Spend ÷ impressions, 33-day flight window. |
| 30% repeat rate · 1.24x annual halo · 4–5 day consideration | Jan–Apr 2026 flight study, same store | Measured repeat behavior; the LTV adjustment is the store's own data, not an industry assumption. |
| 3x projection (FLOOR $98,554 / BASE $157,500 / UPSIDE $163,500 per wk) | Computed 2026-11-05 from the measured inputs above | Measured lift rate × 3x spend × scenario multipliers; every assumption labeled in Section 06. |
(1) The lift band is real and disclosed: $146k–$206k over the flight, depending on model. (2) Eight media states finished below market pace — offsets on the table, not footnotes (CO, HI, MT, NE, OR, PA, VT, WA). (3) The 70% new-customer share is a labeled estimate from the measured 30% repeat rate; the measured new-customer count lands when the warehouse backfill completes. (4) Attribution (the 1.55x weekly reports) and lift (the holdout study) are different measurements: attribution is what the tracking system sees; lift is what the counterfactual proves. This case uses each where it is valid. (5) The CPM at 3x volume is the one external variable the plan does not control — the FLOOR case prices its movement.
3x the measured engine, pointed at the 24 holdout states, the Southern Belt, and the pure-media markets — and Frøya walks into Black Friday week adding ~1,200 new customers a week on a channel that has already proven it is incremental.