Running since December 2025, Freq 2.16, ROAS still 4.11. Hook rate (17.7%) is modest but this is a very short video — p25 is reached inside 3 seconds, meaning every viewer who stops is essentially committed. The creative pre-qualifies buyers before the click: low CTR (1.39%) = only genuinely interested people click, and they buy. The formula: very short + clear product moment = high-intent audience. Don't pause this. It's the account's benchmark.
3s17.7% hook
p25~100% of 3s* (short video)
Hold34.7%
#2
V1527 + CW700 | 26 Nov 2025 Creator Whitelist
The blueprint — best retention curve in the account
This is the template every new creative should chase. 31.2% hook — good, not spectacular. But 77.8% of people who stop are still watching at p25. Most ads lose 60–80% of viewers between 3s and p25. This one loses 22%. The content earns attention in the first second and holds it. Running from the creator's account removes the "ad" feel. This is the retention formula: moderate hook, exceptional body. Nov 2025, still delivering at ROAS 2.82.
Another very short video — p25 is reached in under 3 seconds. 56.5% of viewers watch to the end. CTR 0.59% and CPC RM1.23 signals high-intent clickers only — people who click have decided. ROAS 3.09 confirms it. This is the warm-audience retargeting format: low volume of clicks, high conversion rate per click. Don't expect volume from this — expect quality. Add a CTA overlay to the end card to push more of those high-intent viewers to act.
A static image consistently outperforming most videos. Low CTR (1.24%) and high CPC (RM0.94) = the image pre-filters. Only buyers click. Contrast this with P1287 (same spend tier, ROAS 1.09, CTR 0.80%): same format, radically worse results. The difference is in the copy angle or image choice — P1318 is doing something that converts, P1287 isn't. Study this difference and make it your static brief template.
60% of everyone who sees this stops scrolling. CTR 7.34%, CPC RM0.13. But ROAS is 2.37 — below V1527 (ROAS 2.82 at RM0.60 CPC). The problem: p25/3s is only 15.9% — after the hook, 84% of viewers bail by the first quarter. They click on curiosity, not conviction. They arrive at the product page unprepared to buy. The fix is not the creative — it's the landing page. Build a dedicated page that mirrors the hook energy of this ad and re-routes this traffic there.
3s60.0% hook ← extraordinary
p2515.9% of 3s ← CRASH
p508.7%
#6
V2559 @lilthhhhhh | 7 May 2026 Influencer
Volume driver — 29 purchases, ROAS 2.21, body needs tightening
Most purchases (29), most spend (RM894). p25/3s is 91% — almost everyone who starts, watches the first quarter. Then the drop sharpens: p50 61%, p75 45%. The influencer's personal delivery holds attention in the first 25%, then the video loses momentum in the second half. ROAS 2.21 is below the 2.5 scale threshold (down from 2.47 all-time). Don't scale yet. Same creator, same hook, tighter edit of the second half.
Underperformers — Where Each Creative Breaks
KILL
P1328 | Carousel General | 20 May — breaks at scroll-stop
Weak hook RM242 spend · ROAS 0.59 · 1 purchase · CTR 0.58% · CPC RM1.87
CTR 0.58% means the carousel format isn't stopping anyone. Of the few who click, almost no one buys (ROAS 0.59). No 3s video views because it's static carousel — zero engagement signal. Kill it and move the RM242 to P1318.
RETIRE
@qisssyy__ | 5 May — breaks at hook AND hold
Weak hookWeak hold RM183 spend · ROAS 0.90 · 1 purchase · Hook 10.2% · Hold 14.1% · CTR 0.44%
Hook rate 10.2% — 1 in 10 people stop. Of those, 14.1% finish. Compare @lilthhhhhh (same format, same week): hook 20.8%, hold 30.3%. @qisssyy__ is generating RM183 spend with CTR 0.44% and 1 purchase. The creator hasn't been briefed well or the audience mismatch is real. No signal to save — retire.
WATCH
P1287 | 16 April — breaks at click quality
Weak offer RM318 spend · ROAS 1.09 · 5 purchases · CTR 0.80% · CPC RM0.87
Nearly the same spend as P1318, but CPP RM63.70 vs P1318's RM27.00. The static isn't filtering correctly — low CTR (0.80%) means the image isn't creating desire, and ROAS 1.09 confirms those who do click aren't committed buyers. Retire and study what P1318 does differently.
WATCH
V2120 | 05 Mar — breaks post-click, not creative
Post-click RM347 spend · ROAS 1.73 · 16 purchases · Hook 40.1% · p25/3s 25.5% · CTR 4.09% · CPC RM0.37
Hook 40.1%, CTR 4.09%, CPC RM0.37 — the creative works perfectly. Video body drops (p25/3s 25.5%) but people are clicking before they've finished watching. The impulse-click audience arrives on the product page and doesn't convert at full value. This is not a creative problem. Fix the landing page for this high-CTR, low-consideration traffic.
Format & Angle Patterns — Malaysia Pixel
Short video (<15s)
WINNER
V1778 ROAS 4.11. V2412 ROAS 3.09. p25 > 3s = viewer commitment from first frame.
Make more short videos. Target under 12 seconds. No narration required — product action is the hook.
Creator whitelist (CW)
WINNER
V1527+CW700: best CPP (RM24.66), best p25 retention (77.8%). Removes "ad" feel.
Scale V1527+CW700. Brief 2 more creator partnerships with the same retention-first format.
Static single image
SPLIT
P1318 ROAS 2.73 (CPP RM27). P1287 ROAS 1.09 (CPP RM64). Same format, 2.4× difference in efficiency.
P1318 is the template. Retire P1287 and study what separates them — likely the copy angle.
Hook-and-bail video
ATTENTION
V2593: 60% hook, 7.34% CTR, ROAS 2.37. V2120: 40% hook, 4.09% CTR, ROAS 1.73. Post-click is the gap.
Creative is fine. Build dedicated landing pages for this impulse-click traffic.
Influencer UGC
SPLIT
@lilthhhhhh: ROAS 2.21, 29 purchases. @qisssyy__: ROAS 0.90, 1 purchase. Same format, different briefs.
@lilthhhhhh: iterate tighter second half. @qisssyy__: retire or completely re-brief.
Carousel
WEAK
P1328: ROAS 0.59, CTR 0.58%. Worst format in account.
Exit carousel entirely. Replace with single image or short video.
The retention formula that drives ROAS
The creatives with the highest ROAS are not the ones with the highest hook rates. They're the ones where p25/3s is above 70%. V1527+CW700: 78% p25, ROAS 2.82. V2412: short video ~100% p25, ROAS 3.09. V1778: short video, ROAS 4.11. Sky-high hooks (V2593: 60% hook, V2120: 40% hook) generate cheap clicks but mediocre ROAS because the viewer clicked on curiosity, not conviction. Brief rule for next batch: the video must earn attention in the first 3 seconds AND hold it through the first quarter. If you can only do one, do the second.
Refresh Priority — Malaysia Pixel
Creative
Format
Action
Rationale
V1527 + CW700 | Nov 2025
CW
SCALE
ROAS 2.82, CPP RM24.66, p25/3s 78%. The blueprint. +30% budget.
V1778 | Dec 2025
Short
SCALE
ROAS 4.11. Evergreen. Don't pause what's working.
P1318 | 11 May 2026
Static
SCALE
Best static. ROAS 2.73, CPP RM27. Scale and use as template for next static brief.
V2559 @lilthhhhhh
Influencer
ITERATE
29 purchases but ROAS 2.21. Body drops at p50. Tighter second half — same creator, same hook.
V2593 | 27 May 2026
Video
ITERATE
60% hook is gold. Build a dedicated landing page for this traffic. Don't touch the ad.
V2412 | 28 April 2026
Short
ITERATE
ROAS 3.09, hold 56.5%. Add CTA overlay to end card — turn watchers into clickers.
V2120 | 05 Mar 2026
Video
ITERATE
40% hook, 4.09% CTR. Creative works — fix the landing page.
P1328 | Carousel General
Carousel
RETIRE
ROAS 0.59. Kill now.
@qisssyy__ | 5 May
Influencer
RETIRE
Hook 10.2%, hold 14.1%, ROAS 0.90. No signal to save.
Best CTR in CPAS account. Only video — 1 in 6 stop scrolling vs 1 in 11 for statics.
P1177 | 09 Mar 2026
Static
RM518
1.12%
RM1.24
2.12
—
—
Best static CTR. Solid spend, manageable CPC, Freq still healthy.
P1047 | 13 Jan 2026
Static
RM215
0.64%
RM0.48
1.35
—
—
Cheapest clicks in account (RM0.48). Low CTR but very low cost — wide-reach prospecting.
LB | BM | Direct (×2)
Static
RM4,102
1.15–1.29%
RM0.80–0.88
5.53–5.61
—
—
FATIGUE — third week flagged, still not refreshed. RM4,100/week burning to an exhausted audience.
All Products | Dynamic Carousel
Carousel
RM1,678
0.73%
RM3.02
3.07
—
—
CPC spiked RM0.54 → RM3.02 in one week. Freq 3.07. Audience exhaustion in real-time. Pause.
P956 | 03 Dec 2025
Static
RM364
0.77%
RM0.77
3.66
—
—
Freq 3.66 — fatigued, CTR and CPC declining.
P935 + CW761M | Nov 2025
CW Static
RM187
0.52%
RM4.68
1.73
—
—
Worst CPC in account (RM4.68). Creator-whitelisted static that is not working.
CPAS has 25 active ads and only 1 is a video — make that 3 this week
V1871 (video, CTR 1.69%) outperforms every static in the CPAS account. The account is almost entirely product photos. The LB|BM|Direct campaigns have been flagged for three consecutive weeks and remain live at Freq 5.5+. The Dynamic Carousel CPC jumped from RM0.54 to RM3.02 in a single week — that's what audience exhaustion looks like in the cost data. Brief 2 new CPAS videos now.
Refresh Priority — CPAS
Creative
Action
Rationale
V1871 | 13 Jan 2026
SCALE
Best CTR (1.69%), only video. Increase budget. Brief 2 new CPAS videos urgently.
P1177 | 09 Mar 2026
ITERATE
Best static CTR. Test a version with a different product angle or headline.
LB | BM | Direct (×2)
RETIRE
Freq 5.5+, third week flagged. Pause both today.
All Products | Dynamic Carousel
RETIRE
CPC RM3.02, Freq 3.07 — CPM inflation is live. Pause and refresh product images.
P935 + CW761M
RETIRE
CPC RM4.68 — worst in account.
Singapore — HYGR.SG
30 May – 5 June 2026 · Account 25376284362047670 · ROAS pixel-tracked
Top Performers
#1
V006-LB-colourvariants – Copy | 12 Apr 2026 Video
ROAS 4.46 — best CPP across all accounts this week at SGD6.96
SGD6.96 cost per purchase — the most efficient buying in the entire company this week. The "colour variants" angle (showing all product colours) answers a purchase decision question before the viewer even clicks: "which one should I get?" 35% hook, 89% p25 retention, 46% hold. Everything works at every stage of the funnel. This is SG's equivalent of V1527+CW700 in MY — it should be running at maximum budget.
3s35.0%
p2589.0% of 3s
p5063.9%
p7553.7%
p9546.3%
#2
SV010-LB-blackpink | 25 Apr 2026 Short Video
ROAS 3.95, thin data — but p25/3s 95% is extraordinary
Two purchases on SGD13 is too thin for ROAS conclusions. But p25/3s of 95% — essentially everyone who starts the video watches a quarter of it — cannot be ignored. The "blackpink" colourway video is very short (same p25 > 3s pattern as V1778 in MY). Identical profile to MY's best ROAS videos. Give it SGD200 immediately and check ROAS in 3 days.
#3
V1527-LB-bamboountilidie | 4 May 2026 UGC Advocacy
Creative is excellent — ROAS crashed to 0.88 this week. This is a post-click problem.
Hook 40.7%, hold 42.9%, p25/3s 79%, CTR 5.05% — every creative metric is excellent. All-time ROAS was 2.55. This week it crashed to 0.88. Both SG placements (IG and FB) show equal ROAS — ruling out a placement issue. The drop is cross-platform, which means the problem is post-click: landing page, checkout, pricing, or a competing Shopee/Lazada promotion this week. Do not change the creative. Freeze budget increases until ROAS recovers.
#4
V003-Deo-challengetosolve – Copy Video
56.9% hook, 5.83% CTR — problem-agitation format is stopping scrollers
Same hook-and-bail pattern as V2593 in MY: extraordinary hook (57%), video body crashes (p25/3s 25.5%), but CTR 5.83% at SGD0.31. "Challenge to solve" is a problem-agitation opening — clearly stopping scrollers across both SG and MY. Only 1 purchase on thin spend. Invest SGD100 to get ROAS data. If the body can be tightened to match the hook energy, this becomes SG's attention-grabbing driver.
Underperformers — Singapore
WATCH
P1080-Deorollon — breaks at conversion
Post-click SGD60 spend · ROAS 0.99 · 2 purchases · CTR 3.47% · CPC SGD0.41
CTR 3.47% and CPC SGD0.41 are decent — people are clicking. ROAS 0.99 says they're not buying at adequate value. The roll-on product static gets clicks but the product page isn't closing. Landing page fix for this product before scaling.
THIN
V001-Deo-PopUpReview – Copy — breaks after hook
Weak body SGD17 spend · Hook 38.7% · p25/3s 13.5% · CTR 5.16% · CPC SGD0.27
38.7% hook, 5.16% CTR, CPC SGD0.27 — top funnel is excellent. p25/3s is 13.5% — the body loses 86% of viewers after the first frame. The "pop-up review" format hooks, then the video doesn't sustain. Tighten the body before giving it more budget.
Refresh Priority — Singapore
Creative
Action
Rationale
V006-LB-colourvariants Copy
SCALE
ROAS 4.46, CPP SGD6.96. Scale to SGD300/day immediately.
SV010-LB-blackpink
SCALE
p25/3s 95%. Give it SGD200 to confirm ROAS signal.
V003-Deo-challengetosolve Copy
ITERATE
Hook 57%, CTR 5.83%. Tighten the body. SGD100 test first.
V1527-LB-bamboountilidie
HOLD
Creative is excellent. ROAS crash is post-click. Investigate before scaling or retiring.
61.5% hook rate. 14.52% CTR at USD0.42 CPC — extraordinary engagement. 53.9% hold — people are watching more than half the video. Then 2 purchases on USD117. ROAS 0.56. The creative is not the problem. A US buyer who watches 54% of a video and clicks is genuinely interested. They're landing on a page built for Malaysian shoppers: no US trust signals, unclear USD pricing, no local reviews, unknown brand. The campaign is correctly paused. Do not make new US ads until the landing page is rebuilt.
3s61.5% hook ← exceptional
p2554.0% of 3s
Hold53.9%
US landing page checklist before relaunch
1. US-based customer reviews (not MY/SG reviews) ·
2. Clear USD pricing, no ambiguity ·
3. US shipping time + cost stated upfront ·
4. Trust signal: money-back guarantee or dermatologist-tested ·
5. Exclude Audience Network (39.85% CTR = invalid traffic) ·
6. Facebook-only placement at relaunch (Facebook showed ROAS 1.12 vs Instagram N/A)
Cross-Country Creative Insights
What Travels vs. What Stays Local
Creative Element
→ SG
→ USA
Why / Why Not
V1527+CW700 retention formula (31% hook, 78% p25)
YES
YES
SG runs its version (bamboountilidie). USA needs a founder-facing version. Format is universal.
Short video <15s, no narration
YES
YES
V1778 (ROAS 4.11 MY), SV010-blackpink (ROAS 3.95 SG) — same pattern, different markets.
Colour variant showcase (V006-colourvariants)
ALREADY
YES
SG invented this at ROAS 4.46. MY hasn't tested it yet. USA audience is visual — strong fit.
Problem-agitation hook ("challenge to solve")
YES
YES
Deodorant is a universal problem. Hook 57% in SG. Universal — just localise the specific problem.
Founder / behind-the-scenes content
YES
YES
HYGR's stated brand voice. Underused in every market. Highest priority for USA where brand trust is zero.
Malaysian influencer UGC (@handles)
PARTIAL
NO
SG audience is different. US audience needs US creators or founder voice.
CPAS catalogue static
NO
NO
Marketplace format. Only works inside Shopee/Lazada ecosystem.
8 New Creative Concepts — Next Batch
01
"Which one is yours?" — slow reveal of every HYGR variant
MY primaryUSAShort video, 8–12sColour variant showcase
Builds on: V006-LB-colourvariants (SG, ROAS 4.46, CPP SGD6.96). Malaysia hasn't run this format. Flat lay of all variants, slow left-to-right pan, text: "Which one is yours?" — no voiceover. This answers the purchase decision before the click. Short video = p25 in 3s = everyone who starts commits. HYGR voice: founder making product choices personal.
02
"Hot. Humid. 3pm in KL. Let's see." — climate stress test
MY onlyShort video, 10–15sProblem-agitation → proof
Builds on: V2120's 40.1% hook (pattern interrupt) + V1527's retention body. Creator steps out at 3pm, applies product, camera checks at 6pm. No narration — the result is the creative. Malaysian heat is a specific, shared experience. HYGR voice: doing real things, not polished DTC advertising.
03
"I've been using this every day for 30 days. Here's what actually changed."
MYSGCreator video, 20–30sBrand loyalty advocacy
Builds on: V1527+CW700 retention blueprint (78% p25). "30 days" creates a story arc — something to stay for. Hook delivered direct-to-camera, no cuts in the first 5 seconds. The retention formula from V1527: commit to the camera, don't look away, don't cut. HYGR voice: behind-the-scenes of real product commitment.
04
"Day 1 in a new country. Everything went wrong. This didn't."
USA primaryFounder video, 25–35sBrand origin story
Builds on: USV013's 61.5% hook + 14.5% CTR proves the US audience is interested in HYGR. They click — they just land on a page with no brand story. This ad delivers it. Founder to camera, no B-roll needed. The US buyer who clicked USV013 wants to know who makes this. Answer that in the ad. HYGR voice: doing cool things in new places.
05
"We got a dermatologist to test it. She wasn't our target customer."
Builds on: V2434-Deo-Dr.Review in SG (50.1% hook on thin spend). Expert reactions surprise people. "She wasn't our target customer" creates intrigue. Film raw, handheld. The tension between expert credibility and HYGR's unconventional voice is the story. HYGR voice: associating with credible people outside the usual circle.
06
Static: one specific review line + product image — no other copy
Builds on: P1318 (ROAS 2.73) vs P1287 (ROAS 1.09) — same static format, radically different results. The hypothesis: the best statics have copy that pre-filters buyers. Use the most specific unusual review: "I wore this on a 14-hour flight" beats "great product!" Product as hero, one review line as the only text. HYGR voice: real people, specific experiences.
07
"Can it last NS?" — National Service stress test
SG onlyShort video, 12–15sLocal culture hook
Builds on: SV010-blackpink (p25/3s 95%) — SG responds to very short, direct videos. National Service is a universal shared experience for Singaporean men. "Can it last NS?" is a specific credible challenge that doesn't need explanation. Show the use case in 12 seconds. Text overlay only, no voiceover. HYGR voice: associating with the toughest normal experience in SG culture.
08
"My honest 90-day review — month 1 I almost quit"
MY primarySGUGC/creator video, 30–45sAuthentic long-term review
Builds on: V1778 (Dec 2025, ROAS 4.11) — still running after 6 months. The audience rewards authenticity over time. "Month 1 I almost quit" creates tension that holds people through the video, addressing the p25-to-p50 drop that kills most long videos. HYGR voice: honest, behind-the-scenes, real experience. 90-day credibility that a 7-day review can't match.
Single Most Important Creative Insight This Week
The formula isn't hook rate — it's what happens after the hook
Every account, every market, same finding this week: the highest ROAS creatives are not the ones with the highest hook rates. They're the ones where p25/3s retention is above 70%. V1527+CW700: 78% p25, ROAS 2.82. V2412: short video, ~100% p25, ROAS 3.09. V1778: short video, ROAS 4.11. V006-colourvariants SG: 89% p25, ROAS 4.46.
The hook-and-bail pattern (V2593: 60% hook, 15.9% p25, ROAS 2.37; USV013: 61.5% hook, 54% p25, ROAS 0.56) generates cheap clicks from curious people who weren't informed enough to buy. Write the brief for next batch with one rule: the video must hold attention from 3 seconds to 25% of duration. If the hook is the only thing you nail, you'll get CTR and not conversions.