← ARCHIVE
TO NAV
CROSSROPE · AUGUST 2026 · ROSEWATER GROWTH · PERCENTAGE PLAYBOOK

Creative Demand
Model.

An output floor of 150 launched ads for August, converted into a mix: what share of the month goes to each persona, awareness stage, format and dated moment. The volume number is the least interesting thing on these pages — the proportions are the plan.

READ ME FIRST

150 is a floor, not a ceiling. Every target below is a % of total output, with ad counts shown at the floor for reference — scale them to whatever you actually ship. Seventeen ads launched through 8/5, so 133 remain across the last 25 days. Every lane below carries the spend and creative count behind it — if a signal has under $2.5K or fewer than three creatives supporting it, it is a capped probe, not a plan.

Output floor
150 LAUNCHED ADS · 133 REMAINING
Architecture
8 EVERGREEN LANES · 5 MOMENT LANES · 6 MICRO
Sources
MOTION 90D · SUPERMETRICS AD-LEVEL · AUG CALENDAR
Executive Summary

The mix, in three numbers.

August runs the two biggest sale windows of the quarter — Back to Fitness (8/8–8/17) and Labor Day (8/27–9/1) — with a product drop between them. But the month's largest opportunity is not a new idea: the account already contains ten proven assets and five expensive ones, and the gap between them is worth more than anything on the roadmap. Every ad is a Moment or it is Evergreen.

The prize · at flat spend
+$32K
Ten assets run 2.96x. Five run 1.97x. Closing that gap is worth ~$32K of revenue on the same budget — 2.35x → 2.66x. No new concept comes close.
Output on proven ground
86% / 14%
129 ads on patterns with real money behind them; 21 capped probes. A signal needs ≥$2.5K and ≥3 creatives to earn share.
Moments vs Evergreen
28% / 72%
42 moment ads, 108 always-on. Back to Fitness opens in two days, so its share moved to Labor Day — the window that can still be built.

The big bet for August — and what changed

Stretch what is already working before inventing anything. Ten assets carry 28.2% of spend at 2.96x; five carry 31.5% at 1.97x. Moving the second group's dollars to the first group's efficiency is worth roughly $32K of incremental revenue at flat spend and takes the account from 2.35x to 2.66x.

This is a revision. An earlier cut of this model put 20% of output behind a 50+ messaging lane and twelve ads behind a News-Style format. Held to the evidence bar, those bets do not survive: Bone Health messaging returns 1.53x, creative tagged Women Over 50 returns 2.68x against a 2.79x blended, and the fourteen ads showing 3.5x+ ROAS average $584 of spend each — two or three purchases apiece. The demographic finding survives because it rests on $42K, not $234: women 45+ are 42% of spend at 2.65x. But that is a casting instruction, not a new lane. She gets cast into the formats that already work.

The Evidence Bar · Before any new idea

Ten assets to stretch. Five to fix.

Every active ad with meaningful spend, sorted into two pools against the 2.35x account benchmark. This is the highest-value page in the deck and it contains no new ideas at all — just the gap between what the account already does well and what it already does badly.

Stretch pool · 10 assets
2.96x
$28,652 — 28.2% of the account. Proven at real scale. More cuts, more hooks, more variants of exactly these.
Drag pool · 5 assets
1.97x
$32,000 — 31.5% of the account. Below benchmark at scale. Seed successors, then retire.
The prize · flat spend
+$32K
2.35x → 2.66x. Worth more than every net-new concept in this deck combined.
▲ STRETCH — ≥$1K SPEND, BEATING 2.35x
07/08/2026_CR_AMP Vibe RopelessNewest big winner, launched 7/8. Ropeless + product news.2.55x$5,9105.8% OF ACCT
Video | Ropeless - Nicole L | Home WorkoutBest big performer in the account — and it is from 2025. 36.3% thumbstop.3.00x$5,5375.4% OF ACCT
Video | JRD Suck At JumpingCardio-hater lane. Low thumbstop (11.1%) but converts who it keeps.2.87x$5,1345.1% OF ACCT
05/15/2026_MemorialDayAd3_CR_Memorial Day SaleThe offer-static template that works. Build Labor Day on this.3.37x$2,4232.4% OF ACCT
Video | JRD Slow vs QuickComparison/contrast structure. Highest ROAS of any asset over $2K.3.49x$2,2712.2% OF ACCT
07/16/2026_Whitelisting_@earthandowlInsense. The sister cut of the same creator ran 2.30x.3.11x$2,0522.0% OF ACCT
Video | Joseph G Transformation | Success StoriesNamed-person success story. Highest thumbstop in the account (41.0%).3.47x$1,6761.6% OF ACCT
06/11/2026_Claire SanFran Fitness Whitelisting 1_RopelessInsense × Ropeless. Held spend across six variants.2.57x$1,5581.5% OF ACCT
05/15/2026_Nicole-MemorialDay2026Proven creator × proven offer template. The whole thesis in one asset.3.02x$1,0631.0% OF ACCT
05/28/2026_CR_Brooke-4RW_CR_SCULPTSculpt system, evergreen, underfed.2.88x$1,0281.0% OF ACCT
▼ DRAG — ≥$2.5K SPEND, BELOW 2.35x
05/18/2026_CR_JRD-Ropeless-Who They're For14.4% of the account. The single biggest efficiency leak in the plan.1.77x$14,63314.4% OF ACCT
Video | JRD Losing Weight | July258.4% of the account, below benchmark, weight-loss angle (2.29x tag).2.20x$8,5598.4% OF ACCT
07/14/2026_Summer Clearance SummerSaleAd3_StaticOffer static — but the wrong skeleton. Memorial Day's equivalent ran 3.37x.2.03x$3,4013.3% OF ACCT
07/16/2026_Whitelisting_@earthandowl (cut 2)Same creator as a 3.11x winner. The delta is the cut, not the talent.2.30x$2,8832.8% OF ACCT
06/09/2026_FathersDayAd3_Sale_StaticSecond underperforming offer static. Confirms the pattern.1.86x$2,5242.5% OF ACCT

Why this page comes before the persona split

The account does not have an idea problem. It has ten assets that work, five that do not, and no systematic process for turning the first group into more of itself. Note what the pools are not split by: JRD-sourced content appears in both — Suck At Jumping at 2.87x and Slow vs Quick at 3.49x sit in the stretch pool while Ropeless-Who-They're-For at 1.77x is the single largest drag. The same is true of @earthandowl, whose two cuts run 3.11x and 2.30x. Grade assets, not sources. That is also why the vendor view on this deck is deliberately thin: two lanes, because the vendor lens was never the one that explains this account.

Creative Volume · 2026 Run-Rate

Volume this year — and the August floor.

Unique ads launched per month, counted by first appearance of a distinct ad name in Meta. The regime changed in May when the testing engine turned on: the account went from ~11 launches a month to 189 in July. All figures are launched ads, one unit across the whole deck.

May–Jul average
112/mo
The trailing run-rate since the engine turned on.
July actual
189/mo
The peak — but internal launches adopted just $124 of spend each.
August floor
150
17 launched through 8/5 · 133 to go. Above the trailing average, below July's peak.

The read

Three months ago this account launched eleven ads a month; last month it launched 189. That is a real capability change and it should not be given back. But July's peak came with an adoption problem: internal production launched 162 ads that adopted $124 of spend each, while Insense launched 26 that adopted $318 each — 2.6x the rate. Volume is not the constraint; concentration of effort is. 150 is the right floor: comfortably above the 112 trailing average, deliberately below the 189 peak, with the difference spent on more cuts of concepts that clear the bar rather than more one-cut concepts that do not.
† AUG = the 150 floor. Launch month = first month a distinct ad name recorded spend. Spend line = Meta $K; August omitted (partial month).

The Persona Architecture · Built from the account's own data

Six core lanes, ranked by evidence.

Crossrope has no formal persona research on file, so this architecture is derived from what the account already knows. Lanes are ordered by how much money sits behind the signal, not by how good the ROAS looks — a 4x on $263 does not outrank a 3.61x on $8.6K.

Core lane · who is insideWhere it sits todayAug output
Ropeless / Small-Space Home ExerciserThe strongest evidence in the account and the home of its best big asset. Sub-lanes: the apartment jumper, the joint-limited restart, the travel/hotel set, the Ropeless convert, the downstairs-neighbour problem.3.61x · $8.6K · n=5TODAY IN PAID18%OF AUG OUTPUT
The Cardio-Hater Cross-TrainerTwo stretch-pool assets already live here. Replace-running, cross-training, footwork, off-day conditioning — and where men actually convert.2.66–3.69x · $9.7K · n=3–7TODAY IN PAID14%OF AUG OUTPUT
The Transformation SeekerOver-funded relative to return at format level, but the named-person execution runs 3.47x. Share holds, bar rises.21.6% of spend · 2.60xTODAY IN PAID10%OF AUG OUTPUT
The Time-Starved ParentAbove blended on a thinner base. Back-to-school timing earns it August.4.2% of spend · 3.20xTODAY IN PAID10%OF AUG OUTPUT
The Beginner / First-TimerProven signal with no dedicated lane today. The objection that blocks the purchase.3.66x · Learn A Skill 3.30xTODAY IN PAID8%OF AUG OUTPUT
The 50+ RebuilderA casting finding, not a messaging one. Capped probe until purpose-built creative proves it beats simply casting her in what already works.42% of spend @ 2.65x — but tag 2.68xTODAY IN PAID6%OF AUG OUTPUT

The evidence bar, and why the 50+ lane moved

A lane earns share when ≥$2.5K of 90-day spend and ≥3 distinct creatives support it. Ropeless clears it comfortably (3.61x, $8.6K, five creatives) and takes the lead. The Beginner lane clears it too and had no home at all before this cut. The 50+ Rebuilder does not clear it as a messaging lane — the Women Over 50 tag runs 2.68x on $1.9K against a 2.79x blended, and Bone Health runs 1.53x — so it drops from 20% to a capped 6% probe. The demographic read behind it is untouched and still true; it is expressed as casting inside the proven lanes instead. The remaining 28% sits in the five dated moment lanes.

Flagged assumption

These lanes are inferred from tag performance, not researched. If Crossrope has post-purchase survey or VOC data it should overwrite this slide — that is the evidence that would let the 50+ lane earn share on merit rather than on a demographic inference.

The Spine · How we cut the month

Every ad is a Moment or it is Evergreen.

The first decision on any brief. A Moment ad is built for a dated window and retired when it closes. An Evergreen ad runs always-on against a persona's durable pain. The split is a target proportion of total output.

● Moments · offer & news driven
28% of output
42 of 150 ads. Back to Fitness, Labor Day, the AMP Vibe drop, the navy restock and the partner highlight. Fifteen days of thirty-one carry a dated hook.
● Evergreen · persona driven
72% of output
108 of 150 ads. Always-on against eight lanes — Ropeless, the cardio-hater, transformation, the ten-minute parent, the beginner, the capped 50+ probe, price objection and three open slots.
◂ MOMENTS 28%EVERGREEN 72% ▸

The rule for any brief

If a brief cannot name the dated offer it is for, it is Evergreen — and it should lead with a persona's pain, not a discount. Promo-tagged creative runs 2.90x against 2.71x for no-offer — a real but modest edge, which is exactly why the cap matters: an account that lives on offer creative falls over the day the sale ends.

Moments · Mapped to the August calendar

Five windows — and their share.

Each window's share of the 42 moment ads, weighted to send volume, revenue impact and remaining build time. Taken from the August marketing calendar as supplied on 8/6.

Window · what is in it% of momentsIMG / VID
1Back to Fitness SaleFLAGSHIPAug 8–17. Ten days, eight emails and two SMS. Opens in two days — this kit is finished or it is late, so its share moved to the window that can still be built. Offer-First Banner pattern (3.06x, n=14).35.7%≈15 ADS10 / 5
2Labor Day Sale · 20% offBEST RUNWAYAug 27–Sep 1. Three weeks of build time and a known-good template. Build on the MemorialDayAd3 skeleton (3.37x) — explicitly not the Summer Clearance (2.03x) or Father's Day (1.86x) ones. Approved by 8/26.35.7%≈15 ADS10 / 5
3AMP Vibe · New Colours DropAug 18. Product news the day after BTF closes. The parent asset is already a stretch-pool winner; this is iteration on a proven concept, not a new one.14.3%≈6 ADS4 / 2
4Navy Handles RestockAug 15. Back in stock mid-flight in the BTF window. Scarcity and waitlist language; three disposable statics.7.1%≈3 ADS3 / 0
5Partner HighlightAug 23. Run through Insense in the partner's own handle. Three video assets, no brand-account version.7.1%≈3 ADS0 / 3
Σ All Moments100% 42 ADS27 / 15

Production scheduling read

Back to Fitness opens in two days — that kit is finished or it is late. Its share came down from the last cut for exactly that reason and moved to Labor Day, which has three weeks of runway and the clearest template in the account. Three offer statics tell the whole story: MemorialDayAd3 returned 3.37x, the Summer Clearance static 2.03x, the Father's Day static 1.86x. Build all fifteen Labor Day assets on the Memorial Day structure — and drop proven creators into it, the way Nicole's Memorial Day cut did at 3.02x. Approved by 8/26, not during. The AMP Vibe drop and the restock are small, dated and disposable; Internal Creative owns both. The 8/23 partner feature runs through Insense in-handle.

View 1 of 3 · Persona Mix

Persona % split — the output target.

The share of total output each lane should get. Ad counts at the 150 floor are shown for reference — scale them to your actual volume. Gold = lanes we are leaning into, or dated moment lanes.

Persona / lane% of outputShare# at 150
Ropeless / Small-Space Home ExerciserPROVEN18%
27 ads
The Cardio-Hater Cross-TrainerPROVEN14%
21 ads
The Transformation Seeker · named-person10%
15 ads
The Time-Starved Parent10%
15 ads
The Beginner / First-TimerUNDERBUILT8%
12 ads
The 50+ RebuilderPROBE · CAPPED6%
9 ads
The Value / Price Objection4%
6 ads
Open probe slotsKILL AT $502%
3 ads
MOMENT · Back to FitnessMOMENT10%
15 ads
MOMENT · Labor Day SaleBEST RUNWAY10%
15 ads
MOMENT · AMP Vibe Drop4%
6 ads
MOMENT · Navy Handles Restock2%
3 ads
MOMENT · Partner Highlight2%
3 ads
Total100%150 floor
Persona Rationale · the "why" sheet

Why each lane earns its share.

Every percentage is anchored to Motion's 90-day tag performance, the age-by-gender read, or a dated window on the August calendar. This is the brief-level "why" for the creative team and the vendors.

Ropeless / Small-Space Home ExerciserPROVEN
18%
27 ADS

The best-evidenced lane in the account. Home Exercisers run 3.61x on $8.6K across five distinct creatives — real money, real sample, well clear of the 2.79x Motion blended. It also owns the account's best big asset: Nicole L's Home Workout Ropeless ad at $5.5K and 3.00x, still the top performer eighteen months after launch. Ropeless is simultaneously the product that unlocks the parent, joint-limited and 50+ buyers. August makes it the largest evergreen lane.

Home Exercisers 3.61x on $8.6K · n=5 · Nicole L $5.5K @ 3.00x
The Cardio-Hater Cross-TrainerPROVEN
14%
21 ADS

Two of the account's ten stretch assets live here — JRD Suck At Jumping ($5.1K @ 2.87x) and Slow vs Quick ($2.3K @ 3.49x). Cardio Enthusiasts carry $9.7K at 2.66x and Fitness Enthusiasts 3.69x across seven creatives. The 'replace running' argument is also where the male audience actually converts: men 25–34 at 2.47x against men 35–54 absorbing $28.5K at 2.00x.

JRD Suck 2.87x on $5.1K · Fitness Enthusiasts 3.69x n=7
The Transformation Seeker · named-person
10%
15 ADS

Right lane, wrong execution. Generic Before-and-After is the biggest visual format in the account — 21.6% of spend — and returns a below-blended 2.60x. But Joseph G's named success story runs 3.47x at the highest thumbstop in the account (41.0%), and Brooke's Sculpt cut 2.88x. The share holds; the bar changes. Real named people with real timelines, not split-screens.

Before/After 21.6% @ 2.60x · Joseph G 3.47x @ 41.0% ts
The Time-Starved Parent
10%
15 ADS

Busy Parents run 3.20x across three creatives on $4.0K — above blended, but a thinner base than the two lanes above it, so it gets share rather than the lead. Back-to-school timing makes August its natural month and the ten-minute promise is the whole product. Kathy, Heather and Brooke already carry the voice.

Busy Parents 3.20x on $4.0K · n=3 · back-to-school timing
The Beginner / First-TimerUNDERBUILT
8%
12 ADS

The lane the last model missed. Beginners run 3.66x and the Learn-A-Skill angle 3.30x on $8.3K — both above blended, both with real money behind them, and neither has a dedicated creative lane today. 'I cannot jump rope' is the actual objection standing between a curious buyer and a $200 purchase, and How-To is a 3.15x format across seven creatives. Cheap to produce, and it feeds every other lane.

Beginners 3.66x · Learn A Skill 3.30x on $8.3K · How-To 3.15x n=7
The 50+ RebuilderPROBE · CAPPED
6%
9 ADS

Demoted from 20% — deliberately. The demographic case is strong and stands: women 45+ are 42% of spend at 2.65x and 65+ women buy at $54. But that is a casting finding, not a messaging one. Creative actually tagged Women Over 50 runs 2.68x — at or below the 2.79x blended — and Bone Health messaging returns 1.53x. She is already buying from general creative. Nine ads test whether purpose-built 50+ messaging beats simply casting her in the proven formats.

W45+ 42% of spend @ 2.65x — but 'Women Over 50' creative 2.68x, Bone Health 1.53x
The Value / Price Objection
4%
6 ADS

Price Promotion is a genuinely proven angle — 3.04x across twelve creatives on $13.2K — but it belongs to the moment lanes, which is where twelve of the fifteen offer assets sit. This is the small evergreen residue: Under-$100 entry sets, comparison-vs-gym maths, the founder answer to 'is a $200 rope worth it'.

Price Promotion 3.04x on $13.2K n=12 · mostly lives in Moments
Open probe slotsKILL AT $50
2%
3 ADS

Three unallocated slots for whatever the team wants to try that has no evidence behind it yet — new formats, new creators, new angles. Explicitly ring-fenced so exploration does not leak into the 86% of output that runs on proven ground. Hard kill at $50 if the hook does not clear a 25% thumbstop.

unallocated · hard kill at $50 spend
MOMENT · Back to FitnessMOMENT
10%
15 ADS

The flagship window: 8/8–8/17, eight emails and two SMS. It opens in two days, which caps how much new production can realistically land — so its share comes down from the last cut and moves to Labor Day, where there is actual runway. Offer-First Banner is a 3.06x format across fourteen creatives; build to that pattern.

8/8–8/17 · opens in 2 days · Offer-First Banner 3.06x n=14
MOMENT · Labor Day SaleBEST RUNWAY
10%
15 ADS

The month's biggest recoverable opportunity. 20% off from 8/27, and three weeks of build time. The comp is unusually specific: MemorialDayAd3 returned 3.37x while the Summer Clearance and Father's Day statics ran 2.03x and 1.86x. The template that works is known — use that skeleton, not the recent ones.

8/27–9/1 · MemorialDayAd3 3.37x vs SummerSale 2.03x / FathersDay 1.86x
MOMENT · AMP Vibe Drop
4%
6 ADS

New colours, 8/18. The AMP Vibe Ropeless ad is a stretch-pool asset in its own right ($5.9K @ 2.55x, launched 7/8) and the LE unboxing cut ran 4.55x on $919. A product-news beat that carries the month between two sales without a discount.

8/18 · AMP Vibe Ropeless $5.9K @ 2.55x · LE cut 4.55x
MOMENT · Navy Handles Restock
2%
3 ADS

Back in stock 8/15, mid-flight inside the Back to Fitness window. Scarcity and waitlist language, statics only, disposable by design.

8/15 · scarcity beat inside the BTF window
MOMENT · Partner Highlight
2%
3 ADS

8/23 partner feature, run through Insense in the partner's own handle rather than as brand creative. The Insense lane is the better-adopting of the two vendors and this is its natural moment.

8/23 · Insense in-handle · 2.47x lane blended
The Sub-Persona Engine · Inside the two build lanes

Every brief names a specific human.

A lane name is a planning label, not a brief. These are the named sub-lanes inside the two best-evidenced lanes in the account — the ones carrying 32% of August output between them. One sub-persona per brief, no exceptions.

ROPELESS / SMALL-SPACE · 27 ADS · 3.61x ON $8.6K
The Apartment JumperLow ceilings, thin floors, downstairs neighbours. Ropeless answers this literally and it is the single most repeatable demo in the account — Nicole L's 3.00x ad is exactly this.7ADS
The Joint-Limited RestartPeople With Injuries runs 2.83x on $5.0K across three creatives. Knees, hips, plantar fasciitis, 'I was told not to run.' No rope, no trip, no impact spike.6ADS
The Travel / Hotel-Room SetConvenience & Portability is the most-funded angle at 18.8% and the worst-returning at 2.25x — the lane is right and the execution is tired. Show the packing, not the adjective.5ADS
The Ropeless ConvertFrom a traditional rope to Ropeless: why the switch, what you lose, what you gain. Product-aware, demo-led, feeds the AMP Vibe drop.5ADS
The Small-Space Sceptic'That cannot be a real workout in a 6ft room.' Comparison maths, calorie proof, floor-space overlay. Cheap statics and carousels.4ADS
THE CARDIO-HATER · 21 ADS · 2 STRETCH ASSETS
Replace The Treadmill RunThe JRD Suck At Jumping lane ($5.1K @ 2.87x). Bored runners, bad weather, treadmill fatigue. Direct substitution argument.6ADS
Slow vs Quick · The Contrast CutSlow vs Quick returns 3.49x — the highest ROAS of any asset over $2K in the account. Contrast hooks run 3.26x. Build more of this exact structure.5ADS
The Lifter's Conditioning DayFunctional Strength and off-day conditioning for people who already train. Small tag base, so keep it modest — but the audience overlaps Fitness Enthusiasts at 3.69x.4ADS
Footwork · Boxing & Court SportsAthlete-adjacent credibility. Cheap to shoot, high shareability, feeds the beginner lane too.3ADS
10k Steps Without The WalkThe 10k-steps cut ran 3.98x on $850 — thin but directionally consistent with the substitution argument that already works at scale.3ADS

Compliance guardrail — read before briefing

The joint-limited sub-lane above, and the entire nine-ad 50+ probe, sit close to health claims. Structure/function language only: describe what the activity does, never what it treats, prevents or cures. No claims about osteoporosis, osteopenia, arthritis or any named condition, and no before/after imagery implying a medical outcome. Every ad in those lanes gets a legal pass before launch, and the same rule applies to Insense creator copy — the creator's own words are still our claim.

Micro-Personas · Named lenses for the creative team

Six faces to actually write for.

Cross-cutting lenses that sit underneath the persona lanes. Each one is a person a copywriter can picture, with the signal that justifies it and the format that suits it.

The Downstairs Neighbour
PROVEN · Home Exercisers 3.61x on $8.6K

Third-floor apartment, 8ft ceilings, someone asleep below. The single most repeatable demo in the account and the spine of its best-performing asset.

Demo video + comparison static
The Bored Runner
PROVEN · JRD Suck At Jumping 2.87x on $5.1K

Treadmill fatigue, bad weather, shin splints. Direct substitution — same cardio, ten minutes, no pavement.

UGC video + contrast cut
The Person Who Cannot Jump Rope Yet
UNDERBUILT · Beginners 3.66x · Learn A Skill 3.30x

The objection nobody is answering. Trip, tangle, embarrassment. How-To runs 3.15x across seven creatives — this is the cheapest lane to build and it feeds every other one.

How-To video + carousel
The Told-Not-To-Run Knee
PROVEN · People With Injuries 2.83x on $5.0K

Runners and walkers whose joints quit. Ropeless answers this honestly — no rope, no trip, no impact spike.

UGC video + comparison static
The Ten-Minute Kitchen
SOLID · Busy Parents 3.20x on $4.0K

Before the kids wake, between school runs. Kathy, Heather and Brooke already speak this natively.

POV UGC + GIF
The Named Success Story
PROVEN · Joseph G 3.47x at 41.0% thumbstop

Not a split-screen. A named person, a real timeline, a specific number. The highest thumbstop in the account — and the fix for a Before/After format running 2.60x on 21.6% of spend.

Testimonial video + case-study static
View 2 of 3 · Awareness Mix

Awareness split — where the month sits.

Derived from Motion's hook-tactic and messaging-theme ROAS, mapped onto the funnel. ROAS shown is the 90-day figure for the tag cluster most representative of each stage, against a 2.35x account benchmark.

◂ UNAWAREMOST AWARE ▸
Stage · what lives there% of output90d ROASVerdict
UnawareBroad-tagged creative is 43.7% of spend at 2.38x — the single largest bucket in the account and below benchmark. Every specific audience tag beats it. This is the strongest argument that the persona model is right; it is also why this stage shrinks.8%12 ADS2.38xshrink
Problem AwareQuestion hooks run 3.33x across seven creatives on $11.7K — high volume and above benchmark, the best-evidenced hook tactic in the account. 'Why did the weight stop moving', 'my knees said no to running', 'I cannot jump rope'.22%33 ADS3.33xlean in
Solution AwareThe biggest proven block: How-To 3.15x (n=7, $10.6K) and Demo 2.93x (n=13, $15.8K). Why weighted rope, why Ropeless, why ten minutes beats forty. Also where the beginner lane lives.28%42 ADS3.15xlean in
Product AwareAMP Vibe, Ropeless, the Sculpt system. Lifestyle-Product statics run 3.10x across eight creatives — proven at real scale, unlike the single 3.94x Product-static reading the last cut leaned on.22%33 ADS3.06xhold
Most AwareBack to Fitness and Labor Day. Offer-First Banner is 3.06x across fourteen creatives on $15.2K — the most-evidenced format in the account. Promo 2.90x vs No-Offer 2.71x. August earns the bump; September gives it back.20%30 ADS3.06xhold (offer)
Total100% 150 ADS2.35x acct

Read this honestly

Motion does not tag awareness stage directly — this mapping is inferred from hook tactic, messaging theme and offer type, and format often signals stage more than content does. Treat the percentages as a planning discipline that stops the month collapsing into all-offer or all-top-of-funnel, not as a measured split.

View 3 of 3 · Format Mix

Format by spend, and the target.

Bars show each asset type's share of Motion 90-day spend (steel = video family, coral = image family). The target column is August's planned output. ROAS is only actionable where enough creatives sit behind it — the notes say which readings are load-bearing and which are directional.

Asset type · the call90d spend shareROASAug target
UGC Video · persona-ledThe volume engine and still below the 2.79x blended. Hold share, raise the hook bar
59.3% OF 90D SPEND
2.58x46.0%69 ADS
UGC Mashup / Whitelisted2.88x across six creatives, and the Insense lane adopts 2.6x better per launch. Grow
9.5% OF 90D SPEND
2.88x12.0%18 ADS
Hybrid (UGC + production)Above blended but only two creatives behind it. Hold — needs a bigger base before it earns share
5.6% OF 90D SPEND
2.96x4.0%6 ADS
Lifestyle Static (product + image)The most under-exploited proven asset in the account — 3.14x across fifteen creatives on $15.3K. Grow hard
16.3% OF 90D SPEND
3.14x18.0%27 ADS
Product / Offer Static3.94x, but only three creatives — treat the number as directional. BTF, Labor Day, AMP Vibe and the restock all need them anyway. Make more
3.4% OF 90D SPEND
3.94x8.0%12 ADS
CarouselEffectively unused. Sizing, colourways, the 5-rope system. Test properly, once
— OF 90D SPEND
6.0%9 ADS
GIF / Motion StaticDated, disposable, cheap — built for the two sales and the restock. Moment assets only
— OF 90D SPEND
4.0%6 ADS
Text-only / News-styleCAPPED AT 3 ADS. 3.84x looks spectacular and rests on two creatives and $842 of spend. Directional at best. Probe, do not multiply
0.9% OF 90D SPEND
3.84x2.0%3 ADS
Total2.35x acct100% 150 ADS

The family rebalance — and the number we are not betting on

Video 74.4% → 62% of output. Image family 20.1% → 38%. The load-bearing case is lifestyle statics: 3.14x across fifteen distinct creatives on $15.3K, one of the best-evidenced signals in the account and genuinely underfed. That is what earns the extra share.

What does not: the 3.94x Product-static reading rests on three creatives and the 3.84x text/News-style reading on two creatives and $842 total. Both are kept — the moments need statics anyway — but the News-style lane is capped at three ads rather than the twelve an earlier cut gave it. Carousels and GIFs get 15 ads between them: a real test, once.

The Matrix · Persona × Format

The whole month on one page.

Pick a row and a column and you have a brief. Counts at the 150 floor; rows reconcile to the persona split and columns to the format split. Colour intensity = priority weight.

Persona / lanePrimary awarenessStaticVideoCarouselGIFTotal
Ropeless / Small-SpacePROVENSolution / Product5201127
The Cardio-Hater Cross-TrainerPROVENSolution / Problem317121
The Transformation SeekerProblem / Unaware212115
The Time-Starved ParentProblem / Solution311115
The Beginner / First-TimerUNDERBUILTSolution29112
The 50+ RebuilderPROBEProblem2619
The Value / Price ObjectionProduct / Most Aware3216
Open probe slotsKILL AT $50any213
MOMENT · Back to FitnessMOMENTMost Aware751215
MOMENT · Labor Day SaleMOMENTMost Aware751215
MOMENT · AMP Vibe DropMOMENTProduct Aware3216
MOMENT · Navy Handles RestockMOMENTMost Aware33
MOMENT · Partner HighlightMOMENTSolution Aware33
Σ Total429396150

How to use it

"Ropeless × Video × 20" means twenty videos split across the five sub-lanes on the sub-persona slide — roughly five apartment-jumper, four joint-limited, four travel-set, four Ropeless-convert, three small-space sceptic. Roughly 40% of every cell should be new cuts of an asset that already holds spend, not a net-new concept. The companion Creative Plan workbook expands every cell into named concepts and cuts, one row per launched ad.

Vendor Playbook · How to use this deck

Two lanes. Your mix, and how to forecast from it.

Both production lanes, with 90-day context. 1) find your card 2) read the mix 3) multiply by your committed volume 4) split it across the persona percentages in View 1, and name a sub-persona on every brief. The 150 is a floor — hold the proportions.

Internal Creative
IN-HOUSE · 123 ADS · 82%
$90.4K / 88.9% of 90d spend at 2.33x — but 2.52x once five legacy assets are set aside
YOUR MIX (OF 123)

Ropeless ~24 · Cardio-Hater ~18 · Transformation ~14 · Parent ~13 · Beginner ~11 · 50+ probe ~8 · Value ~6 · probes ~3 · all 39 non-partner moment assets.

POINT IT AT

Iteration before invention. Nine of the ten stretch-pool assets are yours, and so are four of the five drag assets — the lane contains both the best and the worst of the account. The July cohort is healthy (162 launched, 2.42x, six that held spend and beat benchmark), so production quality is not the problem. Adoption is: $124 of spend per launch against Insense's $318. Ship fewer distinct concepts and more cuts of the ones that clear.

THE ONE FIX

Every concept that clears a 25% thumbstop gets three cuts before anything new is greenlit. Kill at $50 if it does not. The @earthandowl pair — 3.11x and 2.30x, same creator — is the whole argument for cut-level iteration.

Insense Whitelisting Partnership Ads
IN-HANDLE · 27 ADS · 18%
$11.3K / 11.1% of spend at 2.47x — and $318 of adopted spend per launch, 2.6x Internal's rate
YOUR MIX (OF 27)

Ropeless ~6 · Cardio-Hater ~5 · Parent ~4 · Transformation ~4 · Beginner ~3 · 50+ casting probe ~2 · Partner Highlight moment 3.

POINT IT AT

The better-adopting lane, and the one with room to grow. Beats benchmark blended (2.47x) and on the July cohort (2.48x), at a cheaper CPA ($67 vs $73). @earthandowl carried $2.1K at 3.11x and Claire SanFran held spend across six variants. Output share goes from 11% of spend to 18% of output. The 8/23 Partner Highlight runs here.

THE ONE FIX

Roster direction, not roster size. The 50+ casting question is cheaper to answer here than anywhere else — sign two creators aged 55+ before Labor Day and let the format do what it already does.

Vendor Scorecard · The monthly grading loop

How each lane graded — 90 days, weighted to last month.

Ad-level Meta data across all 389 active ads, classified by ad-name convention: "Whitelisting" in the name → Insense; everything else → Internal Creative. The grade weighs two lenses: 90-day blended efficiency and the July launch cohort — because a strong blended number can be riding on a few old winners, and a weak one can be dragged by a few old losers. Account blended: 2.35x on $101.6K. "Held" = an ad that took ≥$500 and beat 2.35x.

Lane · feedback90d spend90d ROASLast-mo launchesGrade
Insense Whitelisting Partnership AdsSmall, efficient, and the better converter of launches into spend: $318 adopted per launch against Internal's $124. Above benchmark on both the 90-day blend (2.47x) and the July cohort (2.48x), at a $67 CPA. Feedback: grow the share to 18% of output and point the roster at 55+ and parent creators.
$11,260 · 11.1%
2.47x2.48x · 2 held / 26A-
Internal CreativeThe whole account, effectively. The blended 2.33x sits just under benchmark — but strip the five drag assets and the remaining 346 ads run 2.52x, comfortably above. Recent production is healthy: July's 162 launches adopted $20.1K at 2.42x with six ads holding spend above benchmark. Feedback: the grade is held down by two legacy JRD assets absorbing 23% of the account, not by the work being made now. Fix adoption, retire the drag.
$90,381 · 88.9%
2.33x2.42x · 6 held / 162B+

How the feedback loop works — and why the grades are close

Grade → share → re-grade. A-grades earn budget and volume share next month; B holds; C gets a narrow, template-backed assignment and one month to convert.

With only two vendors the interesting comparison is adoption rate, not blended ROAS. Internal's 2.33x looks like the weaker number until you strip five legacy assets — the remaining 346 ads run 2.52x. The real gap is that Insense turns each launch into $318 of adopted spend against Internal's $124. That is a 2.6x difference in how efficiently a brief becomes budget, and it is why Insense goes from 11% of spend to 18% of output while Internal's assignment is fewer concepts, more cuts.
Read with care: these are in-platform pixel ROAS and will not match the Motion figures used elsewhere — different attribution. Compare lanes to each other, not to the MER plan. ~57 active ads still use pre-convention "Video | …" names, including three of the ten stretch assets.

Where to focus · This month

The eight moves that earn the mix.

The actionable layer — each move names a share, an angle and an owner.

BET · STRETCH THE TEN
Ten assets are 28% of spend at 2.96x
five assets are 32% of spend at 1.97x

This is the month's whole thesis. The account already contains ten proven assets — AMP Vibe, Nicole L's Home Workout, JRD Suck At Jumping, MemorialDayAd3, Slow vs Quick, @earthandowl, Joseph G, Claire SanFran, Nicole's Memorial Day cut and Brooke's Sculpt — running at 2.96x on $28.7K. It also contains five assets running at 1.97x on $32.0K. Moving the drag dollars to the stretch pool's efficiency is worth ~$32K of incremental revenue at flat spend, and takes the account from 2.35x to 2.66x. No new concept in this deck is worth as much as that.

CUT · THE THIN-SIGNAL BETS
Bone Health is 1.53x, not an opportunity
14 ads at 4x+ average $584 of spend each

The previous cut of this model put 20% of output behind a 50+ messaging lane whose flagship evidence was a $234 ad, and twelve ads behind a News-Style format carried by one $391 creative. Under scrutiny the tags do not hold: Bone Health returns 1.53x and Women Over 50 returns 2.68x against a 2.79x blended. Fourteen ads in the account show 3.5x+ ROAS and average $584 of spend — that is two or three purchases each. Those are hypotheses, not findings. The 50+ lane drops to 6% and every thin signal is now a capped probe.

KEEP · THE DEMOGRAPHIC READ
Women 45+ are 42% of spend at 2.65x
that is a casting brief, not a messaging brief

The one part of the old thesis that survives scrutiny, because it rests on $42K of spend rather than $234. Women 55–64 return 2.82x, 65+ return 2.98x at a $54 CPA, and men 35–54 absorb $28.5K at 2.00x. But she is already buying — from general creative. So the correction is who appears in the proven formats, not a separate menopause lane. Cast her into the Ropeless demo, the How-To and the offer statics that already work.

FIX · ONE AD IS 14.4% OF THE ACCOUNT
JRD Ropeless-Who-They're-For · $14.6K @ 1.77x
plus JRD Losing Weight · $8.6K @ 2.20x

Two legacy assets absorb 23% of the account below benchmark and drag the Internal grade from 2.52x to 2.33x on their own. This is not a vendor problem — JRD content also produced two stretch-pool winners at 2.87x and 3.49x. It is an asset-concentration problem. Seed direct successors on the Suck At Jumping and Slow vs Quick skeletons now, before fatigue forces the issue mid-BTF.

BUILD · LABOR DAY ON THE RIGHT SKELETON
MemorialDayAd3 3.37x
SummerSale 2.03x · FathersDay 1.86x

Three offer statics, three very different outcomes, one clear template. Labor Day has three weeks of runway — the only moment this month that does. Build all fifteen assets against the Memorial Day structure, drop the proven creators into it (Nicole's Memorial Day cut ran 3.02x), and have them approved by 8/26.

GROW · THE BETTER-ADOPTING VENDOR
Insense: $318 adopted per launch
Internal: $124 per launch

Insense converts launches into spend 2.6x more efficiently than internal production, beats benchmark on both blended and cohort ROAS, and buys at a cheaper CPA. It is 11% of spend today. Output share goes to 18%. This is also the cheapest way to answer the 50+ casting question — buy the face rather than build the lane.

REBALANCE · TOWARD STATICS THAT WORK
Lifestyle statics 3.14x across 15 creatives
18% of output, up from 16.3% of spend

Not the single 3.94x reading the last cut leaned on — this is $15.3K across fifteen distinct creatives at 3.14x, one of the best-evidenced signals in the account and genuinely underfed. Image family goes to 38% of output. Carousels and GIFs get 15 ads between them: a real test, once, rather than a permanent maybe.

HOUSEKEEPING · NAMING
Rename legacy 'Video | …' assets into the convention
~57 ads have no date prefix

About a sixth of the active library predates the MM/DD/YYYY convention. Three of the ten stretch assets are legacy-named, which means the best-performing creative in the account cannot be graded automatically. Rename or retire this month.

Methodology & Sources

How this model gets built.

The 150 floor was set against the 2026 launch run-rate. The percentage splits come from Motion 90-day tag performance filtered through an evidence bar — ≥$2.5K of spend and ≥3 distinct creatives — then Supermetrics ad-level history for the stretch and drag pools, then the August calendar. Refreshed weekly; targets reset the first Monday of each month.

August inputs
Output floor (confirmed by Max)150 launched ads
Launched through 8/517 ads · 133 remaining
Evidence bar≥ $2.5K spend AND ≥ 3 creatives
Output on proven ground86% · probes capped at 14%
Moments / Evergreen28% / 72%
Video / Image family62% / 38%
90d Meta spend / ROAS$101.6K · 2.35x benchmark
Stretch pool / Drag pool28.2% @ 2.96x / 31.5% @ 1.97x
VendorsInternal Creative 82% · Insense 18%
Vendor grading cadence1st Monday · grades move next month's share
Sources & basis
Motion · 90d SPEND + glossarytag ROAS with spend and creative counts (top 60, $94.2K)
Motion · age × gender 90dthe casting read — 42% of spend is women 45+
Supermetrics FA · 13-month ad-levelvolume chart, launch cohorts by first appearance
Supermetrics FA · 90d daily ad-levelstretch/drag pools · 6,382 rows · $101.6K · 389 ads
August marketing calendarmoment windows (client screenshot, 8/6)
Ad-name convention'Whitelisting' in name → Insense; all else → Internal Creative
Companion Creative Plan150-row workbook · one row per launched ad
CadenceWeekly · Mondays

Flagged assumptions

1. Crossrope has no persona research or post-purchase survey data on file — lanes are inferred from Motion tags and demographic performance, and should be overwritten if research exists. 2. Awareness stage is inferred, not tagged. 3. Launch months are derived from first appearance of an ad name, which slightly understates July. 4. Moment windows are read from the August calendar screenshot supplied 8/6; the Back to Fitness kit is presumed already in production. 5. Vendor classification is mechanical: "Whitelisting" in the ad name → Insense, all else → Internal Creative. Any Insense asset not carrying that token is currently counted as Internal. 6. The $2.5K / 3-creative evidence bar is a judgement call, not a statistical test — it is set where the account's own noise floor sits (fourteen ads show 3.5x+ ROAS on an average of $584 each). Lanes below the bar are not judged wrong, only unproven, and are sized as probes so they can earn share next month.

The one thing to remember

These are proportions, not quotas. 150 is the floor and exceeding it is good — whatever ships, hold the mix, and make every brief name a persona and a sub-persona. The single test of whether August worked: did the 50+ Rebuilder actually get thirty ads?

Rosewater Growth  ·  Crossrope  ·  Aug 2026