Which creatives click, which collect emails, which get paid

Headline

• Two creatives carry the account. Says Who 15s ($547) and Chair workout 32s ($440) are 88% of spend. The other 14 have $1–$60 each and no readable signal.
• Says Who wins every click and lead metric and has never made a sale. 8.6% CTR, $0.27 a click, $2.14 a lead, 17 checkout taps, 0 cards, across five campaigns.
• Chair workout is the only creative that has ever produced money. 2 paid and 3 live trials, all from the purchase-objective campaign, at 3–10× Says Who's cost per click.
• Objective moves the numbers more than creative does. The same Chair ad costs $0.68 a click on the leads objective and $3.64 on purchase, and only the purchase traffic pays: 5 of 8 checkouts completed there, 0 of 24 on every leads-objective ad combined.
• Correction to the Sep 4/5 notes. The Sep 2 trial59 trial came from Overhead guided workouts (clicked Aug 29, converted four days later), not Chair. That woman cancelled on Sep 3: "This should be called 80Queen! The moves are way too simple".

Says Who against Chair workout

All campaigns combined. Bars are scaled within each panel; the number is the value.

Every creative, all campaigns combined

CTR is link clicks over impressions. Leads are PostHog email_captured by quiz id, with the Meta pixel lead count beside them; CPL uses PostHog. The Aug 7–10 validation campaign predates the PostHog ingest fix, so its on-site numbers are missing and its CPL is from the pixel. Checkouts are PostHog checkout_started. Trials and paid are Stripe live. Rows in grey spent under $10 and cannot be read.

Every campaign and creative

The identity-test Chair row combines the 50queen arm ($114.87, the $14.99 four-weekly on Aug 30), the fiftystrong arm ($99.33, the $49 twelve-week on Sep 1 and the Sep 2 monthly trial) and the Sep 4–5 fs.app ad ($116.80, both Sep 5 belly trials). Meta's pixel credits both purchases to the 50queen ad; the ad-set id Meta appends to the landing URL and Stripe agree on one each.

What the numbers say

Says Who is a click and lead machine that does not sell

It has the lowest CPC in every campaign it shares with Chair workout (2–5× cheaper), the highest CTR, and a $1.42–2.12 CPL on the lead objective. It has also produced 17 checkout taps and not one card. Its one purchase-objective run (purchases-launch, $17, 16 entrants) got 2 leads and 0 checkouts, too small to read. The obvious untested combination, Says Who on the purchase objective at real budget, is built and paused: Identity · fs.app · Says Who · 15s · validation cut (120252431918130131).

Chair workout is the only creative with paying customers, and only on the purchase objective

On identity-test it is $331 for 2 paid + 3 trials: $66 per conversion, $166 per paid today, $66 per paid if all three trials bill. The same creative on quiz-leads-v2 produced 24 leads at $2.58 and 7 checkouts, none completed. Its CTR is 3.6–5.0% against Says Who's 5.9–8.1% in the same campaigns, and its cost per click is 2–5× higher.

The leads objective has never produced a card

24 checkout taps across quiz-leads-launch and quiz-leads-v2, 0 completed. The purchase objective: 9 taps, 6 completed. That split holds for both creatives, so it is the audience Meta picks, not the ad.

The purchases-launch long tail is noise

Twelve creatives under $40 each over two days, 0–4 clicks apiece. Three of four statics got zero clicks at $28–59 CPM; rank08 got zero clicks on 56 impressions. Overhead guided workouts got the one late trial from 9 clicks, and she cancelled inside a day with a complaint about difficulty. Abs mirror (3 leads on 7 clicks) is the only one with a hint.

LPV traffic is worthless

12.8% CTR and $0.04 clicks on Says Who turned 406 entrants into 1 lead.

Sample size

Two paying customers and four trials. Click and lead differences between Says Who and Chair workout are large and consistent across five campaigns. The sales difference rests on 2 payments and cannot yet separate "Chair sells" from "Chair happened to run on the objective that sells".

Trials in flight

Paid so far: $14.99 four-weekly (Aug 30, active, auto-ends Nov 22) and $49 twelve-week (Sep 1, one-off).

Recommendations

1. Reset the spend cap now. Nothing has served since Sep 5; every day the trial windows run without new data.
2. Activate Says Who in the purchase set (120252431918130131, already built) alongside Chair at equal budget. That is the single highest-value unknown: the best click economics in the account on the only objective that has ever converted. Success looks like a CPC under $1.50 on the purchase objective with any completed checkout in the first $150.
3. Stop counting leads-objective checkouts as intent and stop scaling that set on CPL. 0 of 24 is the number. Keep it at a trickle only if the email list is worth $2.50 an address.
4. Do not revive the purchases-launch tail. If a third creative is wanted, Abs mirror is the only candidate with a signal; the statics and rank08 are done.
5. Read again on Sep 12 after the Sep 9 and Sep 12 trials resolve. Three live trials decide whether Chair's CAC is $66 or $166.

Sources

• Meta act_1761459111652871/insights?level=ad&date_preset=maximum with actions, inline link clicks and video fields; daily increments for delivery windows; each creative's landing URL for the UTM mapping. Account readback: amount spent $200.00 of a $200.00 cap.
• PostHog prod 559701 Per quiz id: first non-empty utm_campaign and utm_content, distinct screens viewed, email_captured, paywall_viewed, checkout_started, trial_started, purchase_completed; Meta data-centre cities excluded. Stripe quiz ids resolved to campaign and creative by the entrant's first URL.
• Stripe live Checkout sessions, subscriptions and charges, plus subscription retrieval for cancellation details.
• Working note .claude/ads/creative-performance-2026-09-07.md in the fiftyworkout repo.