Ad Creatives Agency · Dubai · UAE · KSA
Ad creatives built on
hook architecture —
tested before scaled.
Most ad budgets scale behind creatives that were never tested. The hook fails in 3 seconds, the retention layer never gets seen, and the budget attribution reports success while the ROAS declines. The Adzyon ad creative system starts with an audit of every active creative's hook performance, builds angle hypotheses from audience psychology, and tests at the velocity algorithmic platforms demand — 48 hours from brief to live test.
+147%
median hook retention improvement when 3-second hook architecture is applied to ad creatives that previously had no documented hook structure — measured against the brand's own baseline 3-second view rate before the creative overhaul
4 angles
distinct creative angles tested per campaign before a primary winning brief is identified, scaled, and set as the new control for the next test cycle — problem-first, proof-first, outcome-first, and offer-first
3 channels
platform-native production across TikTok, Meta, and Google with channel-specific hook timing, format specifications, and text overlay placement — one brief, three native assets, one test cycle
02 / Why Creatives Fail
Hook abandonment. Platform-generic assets. Budget scaled before testing. Three failures, one outcome.
Three structural failures prevent ad creatives from converting at scale: the hook fails before the argument layer is seen (which means the brand's best proof and offer is presented to nobody); the same asset runs on TikTok, Meta, and Google simultaneously (wrong format for every platform at once); and budget is scaled before any variant has reached statistical significance (scaling a variable that hasn't been validated). Each failure is independent. Most programmes have all three.
Hook failure before the first 3 seconds — the creative never gets a chance to convert
Algorithmic platforms allocate attention in milliseconds. A creative that doesn't interrupt the scroll and establish relevance in the first 1–3 seconds is never seen — the viewer skips, the algorithm registers low engagement, and the ad is deprioritised in future auctions. Most brands don't measure 3-second view rate per creative, don't segment it by angle or format, and don't use it as the primary signal for creative iteration. The hook is treated as a header, not as the most leveraged 3 seconds in the entire acquisition funnel.
Consequence
The brand's strongest arguments — proof, outcomes, offer — are presented to an audience that left before the hook resolved. Conversion events are attributed to budget spent behind a creative that failed before the argument layer was reached. ROAS degrades while the creative is still running because no hook performance signal triggered an iteration.
Platform-generic creative — one asset repurposed across every channel
Running the same 30-second landscape brand video on TikTok, Meta, and Google simultaneously is wrong for every platform at once. TikTok requires native vertical 9:16 video with a creator-native tone, front-loaded hook, and sound-on architecture. Meta feed performs best with static images for direct-response objectives. Google requires intention-matched copy tied to the query or audience intent signal, not a brand video. A single horizontal asset letterboxed into a vertical format, with English audio and Arabic subtitles added in post, is not platform-native creative — it is a repurposed asset that signals to the platform algorithm that the advertiser doesn't understand the content environment.
Consequence
Platform-specific algorithms penalise creative that doesn't match the content format and audience behaviour patterns of the placement. The creative loses the auction before targeting decisions matter. The brand pays full CPM rates for creative that the platform's own quality score would discount if the placement were evaluated correctly.
Budget scaled before creative tested — spending against an unknown variable
Increasing ad spend behind an untested creative is scaling an unknown. Without a structured test cycle — documented hypothesis, variant, measurement period, and significance threshold — there is no evidence that the creative is profitable at 2× budget, marginal at 3×, or ROAS-negative at 5×. Most brands scale spend when the first few days of data look positive, before the creative has run long enough to reach statistical significance. The early positive signal is sampling noise, not performance evidence.
Consequence
Budget is scaled on noise rather than signal. When ROAS eventually declines — as creative fatigue sets in or early-sampling luck regresses — the brand has no documented learning about which creative element was responsible. The budget is now committed to a creative that is degrading, and there is no tested replacement ready to scale.
Creative performance benchmarks
3 sec
the window in which the hook must interrupt the scroll, establish relevance, and create a reason to keep watching — before the viewer's pattern recognition decides to skip
3–5×
more ROAS impact from testing hook angle vs. testing design elements — the highest-leverage creative variable is the first 3 seconds, not the button colour or the headline font
<20%
of ad creative tests reach statistical significance before the budget is cut — because creative testing requires a minimum volume threshold per variant that most campaigns never plan for
03 / The Ad Creative System
Audit, brief, produce, test. In that order.
Four stages from creative audit to compounding creative library — each producing the output the next requires. Creative Audit produces the hook performance heatmap: every active creative mapped by 3-second view rate, hold rate, and ROAS — the baseline that makes brief architecture evidence-based rather than instinct-driven. Brief Architecture translates that analysis into testable angle hypotheses before any production begins — one documented hypothesis per creative variant, not one brief for a batch. Production and Testing builds platform-native creative variants from the brief and runs structured tests with a defined significance threshold — tests run until 95% confidence, not until budget is spent. Scale and Iterate enters winners into the scaling protocol and builds the next hypothesis from winner analysis, so each cycle tests against the current best performer rather than the original baseline.
Why the creative audit precedes brief architecture
Most creative programmes start with a brief because it feels productive. The creative audit is slower and less satisfying — it produces a report rather than an asset. But a brief written before the audit is a hypothesis formed without data: the angle chosen is the one that feels most compelling to the team, not the one that the current creative library's performance data indicates is most likely to outperform the control. The audit prevents this. It produces the evidence that makes the brief's hypothesis defensible rather than instinct-driven.
- 01
Creative Audit
Measure the current ad library before forming a single hypothesis. What is the 3-second view rate per creative? What is the hook hold rate — the percentage of viewers who watch past the hook into the argument layer? What is the ROAS per creative, per angle, per format? Where is the largest performance gap in the creative library? The creative audit produces a hook performance heatmap: every active creative mapped by view rate, hold rate, and ROAS — showing which angles are generating acquisition economics and which are spending budget without producing measurable learning. The audit also identifies the angle and format gaps: the creative hypotheses that haven't been tested and the platforms where native creative hasn't been produced.
Output: Creative audit — hook performance heatmap per creative, 3-second view rate per angle, hold rate analysis, ROAS per format, angle and format gap register - 02
Brief Architecture
Build the creative brief from audience psychology, not from brand preference. What is the audience's primary decision-making trigger? Is it problem awareness (they don't know the solution exists), proof urgency (they know the category but haven't committed to a brand), outcome desire (they want the result and are evaluating options), or price sensitivity (they've decided to purchase and are comparing offers)? Each trigger maps to a different hook angle, a different retention architecture, and a different offer framing at the conversion layer. The brief documents: hook format per platform, hook copy (5 variants), retention argument structure, offer framing, CTA specification, and primary measurement event per channel. One brief per angle hypothesis. No creative goes into production without a documented hypothesis.
Output: Creative brief per angle — hook format spec, 5 hook copy variants, retention architecture, offer framing, CTA, primary measurement event, and platform specifications - 03
Production and Testing
Produce platform-native creative variants from the brief and launch structured tests. Three to five variants per hypothesis — each with a single variable changed against the control. TikTok: vertical 9:16 video, sound-on hook architecture, creator-native tone, text overlay for sound-off fallback. Meta: static and video variants tested independently, scroll-stop visual, offer-prominent copy. Google: intention-matching copy and display creative that aligns with the search query or intent signal. Every test is launched with a documented hypothesis, a defined measurement period, a primary metric (3-second view rate at hook stage; ROAS at conversion stage), and a significance threshold before a winner is called. Tests run until 95% confidence is reached — not until budget is spent.
Output: Live test — platform-native creative variants per hypothesis, documented test measurement, significance tracking per channel, preliminary results at 80% confidence - 04
Scale and Iterate
Scale winners, retire underperformers, and build the next hypothesis from winner analysis. The winning creative is not a permanent asset — it is the new control for the next test cycle. What in the winning variant explains its performance? Is it the hook angle, the retention structure, the offer framing, or the CTA? The analysis produces the next brief: a more refined hypothesis built on what the winning creative revealed about the audience's conversion trigger. Over four test cycles, the ROAS improvement compounds — each winner outperforms the previous control by a measurable margin, and the creative library becomes a proprietary asset that competitors cannot replicate by copying surface-level creative elements.
Output: Production deployment — winning variant scaled per channel, underperformers retired, winner analysis brief, next-cycle hypothesis, quarterly ROAS improvement report
Want to see how this applies to your funnel?
A senior strategist reviews your specific setup — complimentary, no pitch deck.
04 / Hook, Retention & Conversion
Every ad creative has three layers. Most brands only design the first.
High-performing ad creative is not a single creative unit — it is three sequential layers, each with a distinct job, a distinct audience psychology requirement, and a distinct primary metric. The hook earns the viewer's attention. The retention layer earns the right to present an offer. The conversion layer converts the qualified viewer. A creative that succeeds at the hook layer but fails at retention is producing qualified impressions with no conversion path. A creative that succeeds at conversion but fails at the hook is presenting an offer to nobody. All three layers must be designed and tested independently.
The scroll-stop moment — 0 to 3 seconds
Interrupt the scroll and establish relevance
The first 1–3 seconds determine whether the viewer stays. The hook must do two things simultaneously: create a visual or auditory interruption that stops the scroll (the pattern-break), and establish that the next 15 seconds are relevant to this specific viewer's situation (the relevance signal). A hook that creates an interruption but fails the relevance test produces high 3-second view rates and low hold rates — the audience paused but immediately decided the creative wasn't for them. A hook that establishes relevance but fails the interruption test is never seen. The most effective hooks combine a direct address ('if you're running Meta ads and your ROAS keeps declining...'), a visual that differs from surrounding content, and a text overlay that communicates the hook for sound-off environments.
What to test
Primary metric: 3-second view rate — the percentage of impressions where the viewer watches past the first 3 seconds. Test variable: hook angle (problem-first vs. proof-first vs. outcome-first vs. offer-first).
Failure signal
Hook abandonment: 3-second view rate below platform average (typically 15–20% for cold audiences). Cause: irrelevant hook angle, slow visual opening, or logo/brand-first frame that signals 'advertisement' before establishing value.
The argument layer — 3 to 15 seconds
Build the case before the offer is presented
The retention layer is where the creative earns the right to present its offer. The audience has committed past the hook — they are now asking: 'Why should I believe this?' and 'What is this actually going to give me?' The retention architecture must answer both questions before the offer is revealed. For cold audiences (problem-aware): the retention layer establishes the problem's real cost, introduces the mechanism (why this product/service solves it), and builds credibility. For warm audiences (solution-aware): the retention layer leads with social proof — testimonial, case study result, or user count — because the problem is already understood. For retargeting audiences: the retention layer leads with urgency or offer differentiation — because they know the brand and need a reason to act now rather than later. The retention architecture is determined by the audience temperature, not by brand preference.
What to test
Primary metric: hook hold rate (25–50% completion rate). Test variable: retention argument structure (problem → mechanism → proof vs. proof → mechanism → outcome).
Failure signal
Premature exit: hold rate below 25% despite a strong hook. Cause: the retention layer doesn't answer the implicit question the hook created, or presents the offer before establishing sufficient credibility for the audience's decision stage.
The offer and CTA — 15 to 30 seconds
Convert the argument into a click
The conversion layer is only reached by viewers who passed both the hook and the retention filter — which means they are the highest-intent audience in the impression set. The conversion layer must present the offer clearly, reduce friction to the click, and make the next step obvious. Three elements determine conversion layer performance: offer architecture (is the CTA asking for the right commitment for this audience's decision stage — trial vs. purchase vs. enquiry?), friction reduction (is the next step presented as low-commitment and easy to take?), and urgency (is there a genuine reason to act now — time-limited offer, Ramadan activation, limited stock — not a generic countdown clock?). The conversion layer is not a place for brand messaging or product features. The argument is complete. The job is to convert the qualified viewer into a click.
What to test
Primary metric: conversion rate and ROAS. Test variable: CTA copy (try free vs. shop now vs. get 20% off), offer architecture (percentage discount vs. AED value vs. free trial), urgency signal (countdown vs. stock vs. none).
Failure signal
Click-to-conversion drop-off: high hold rate with low ROAS. Cause: the offer doesn't match the audience's commitment threshold (asking for a purchase from an audience that was ready for a trial), or the CTA creates friction rather than reducing it.
05 / Creative Angles
Four audience psychology triggers. Four distinct creative angles. One angle per hypothesis.
The creative angle is not the visual style or the copy tone — it is the psychological trigger the hook is designed to activate. Problem-first, proof-first, outcome-first, and offer-first each speak to a different decision-making stage. Matching the wrong angle to the wrong audience stage produces a high-impression creative with no conversion path: a problem-first hook presented to an offer-motivated audience restates a problem they've already decided to solve. The brief development process starts by identifying where in the decision funnel the target audience segment sits — and writing the angle that meets them there.
Angle 01
Problem-first
Cold — problem-aware, solution-unaware
Identifies with a problem but hasn't considered this category as the solution
Problem-first creative names the problem the audience already feels but hasn't articulated as a purchase opportunity. The hook identifies a specific pain point, quantifies its cost, and positions the product as the mechanism that eliminates it. The retention layer establishes credibility through the mechanism — why this approach works when others haven't — before presenting the offer. Problem-first is the highest-performing angle for cold audience acquisition because it meets the audience where they are (problem-aware) rather than where the brand is (solution-aware). It requires research to identify the problem formulation that the audience recognises as their own — not the problem the brand thinks the audience has.
Hook structure
'Most [audience description] are losing [cost/outcome] because [specific problem]. Here's why — and what to do instead.'
Best for
Cold audience acquisition on TikTok and Meta. Top-of-funnel campaigns where brand awareness is low and category awareness is high.
Angle 02
Proof-first
Warm — solution-aware, brand-evaluating
Knows the category, is evaluating this brand against alternatives
Proof-first creative leads with a measurable result from a real customer — ROAS lifted, revenue grown, cost reduced, time saved — before explaining the mechanism or presenting the offer. The hook establishes a specific, credible outcome that the target audience recognises as meaningful. The retention layer then explains how that outcome was produced, which serves as the mechanism introduction for audiences who are already solution-aware. Proof-first consistently outperforms problem-first for warm retargeting audiences and mid-funnel campaigns because the audience has already processed the problem — they need social proof to differentiate one solution from another, not another framing of the problem they already understand.
Hook structure
'[Client/customer name or type] went from [before state] to [after state] in [timeframe]. Here's exactly what we did.'
Best for
Warm retargeting on Meta. Mid-funnel B2B and SaaS campaigns. Any audience that has seen problem-first creative 3+ times and is now evaluation-stage.
Angle 03
Outcome-first
Decision — brand-familiar, weighing commitment
Has evaluated the category and brand; needs a reason to commit now
Outcome-first creative opens with the desired end state — the life or business after the purchase decision is made — rather than the problem or the proof. The hook describes the specific outcome the target audience wants: 'Your ad account is consistently profitable. Every campaign is scaling. You know exactly which creative is working and why.' The retention layer explains the path from current state to that outcome through the product. Outcome-first works for audiences who are already brand-familiar and have decided they want the result — the creative's job is to connect their desired outcome specifically to this product rather than to a competitor alternative.
Hook structure
'Imagine [specific desired outcome in vivid detail]. That's what [product/service] is built to produce. Here's the system.'
Best for
Retargeting audiences who have visited the pricing or product page. Warm B2B audiences who have consumed multiple content pieces. Late-funnel conversion campaigns.
Angle 04
Offer-first
Conversion — purchase-motivated, price-comparing
Has decided to purchase in this category; is evaluating price and offer terms
Offer-first creative leads with the economic value of the transaction — the discount, the free trial, the bundle, the BNPL installment price — before any argument or proof. The hook is the offer: 'AED 299 off — this week only' or 'Start your free 30-day trial — no card required.' The retention layer justifies the offer credibility (this is a real discount; the trial includes full product access) and reduces friction (the next step is one click, takes 2 minutes, requires no commitment). Offer-first is the lowest-funnel angle and performs best for price-motivated audiences who have already decided to purchase in the category and are comparing offer terms. It underperforms for cold audiences who haven't established relevance — a discount on a product they don't understand is not a compelling proposition.
Hook structure
'[Specific offer — AED value, percentage, trial terms] — [urgency signal]. [CTA that reduces commitment friction].'
Best for
Conversion campaigns for existing audiences. Ramadan and Eid activation. Coupon affiliate traffic landing pages. Price-sensitive ecommerce categories with high competitive density.
06 / Creative Testing
A creative testing programme is not a series of experiments. It is a compounding system.
Three disciplines separate a creative testing programme from ad hoc creative production: one variable per test (so the winner analysis produces a learning, not just a result); statistical significance before scaling (so the budget decision is based on evidence, not directional data); and the winner as the new control (so each test cycle builds on the previous cycle's improvement rather than resetting to the original baseline). These three disciplines are what produce a creative library that compounds in performance rather than a library that accumulates assets without improving ROAS.
One variable per test
Only one element changes between control and variant.
A creative test that changes the hook angle, the offer, and the CTA simultaneously produces a result that can't be interpreted. The winner outperformed the control — but which variable explains the win? Without a single variable, the learning is 'this combination was better' rather than 'this specific element explains the performance difference.' One variable per test is not a constraint on creative production speed — it is the discipline that produces actionable learning rather than inconclusive data. Four tests with one variable each produce four learnings. One test with four variables produces one uninterpretable result.
Significance before scaling
A budget decision is not a significance threshold.
The most common creative testing failure is calling a winner when the budget is spent rather than when statistical significance is reached. A variant that is 'trending positive' after 3 days of data has not produced evidence that it will outperform the control consistently — it has produced early sampling data that is indistinguishable from noise in most traffic volumes. The correct significance threshold is 95% confidence level, with a minimum of 100 conversions per variant for conversion-stage tests. Until significance is reached, the test is still running. Cutting the test before significance produces a 'winner' that is statistically no better than the control in 1 in 20 tests — which means 1 in 20 decisions is made from random variation rather than performance signal.
Winner as new control
Each test cycle builds on the previous winner — not the original baseline.
The compounding structure of a creative testing programme is what separates it from a series of one-off experiments. When the winning variant replaces the control at the start of the next test cycle, the programme is testing against the current best performer — not against the original creative. Over four cycles, the creative library reflects four cumulative improvements: each winning element identified and locked, and each new hypothesis testing the next layer of opportunity. A programme that resets to the original control at each cycle is producing the same learning range in cycle 4 as in cycle 1. The compounding structure means cycle 4 is testing refinements to a creative that already incorporates three proven improvements.
07 / Platform-Native Production
TikTok, Meta, and Google require different creative — not the same asset in three aspect ratios.
Platform-native creative is not a resolution change. TikTok requires a creator-native tone and a sound-on hook architecture. Meta requires static images for feed direct-response and vertical video for Stories and Reels. Google requires intention-matched copy and display creative tied to the audience signal that triggered the placement. Producing native creative per platform requires understanding what the audience expects from the content environment — not applying a single brand creative template to three different contexts.
Platform 01
TikTok
Primary metric: 3-second view rate + hook hold rate (25% completion)
TikTok creative must feel native to a content environment where every surrounding video is creator-made, fast-paced, and direct-to-camera. A polished production value that signals 'television advertisement' is penalised by the audience immediately — the pattern recognition identifies it as an interruption, not content, and skips before the hook resolves. TikTok creative works when it adopts creator conventions: direct address to camera, sound-on architecture (the hook must work with audio, not despite the lack of it), text overlays that serve as subtitles rather than headlines, and a pacing that matches the cut-frequency of organic TikTok content. The hook must resolve in 1–2 seconds — TikTok audiences are faster than Meta audiences at the skip decision. BNPL display in the creative (Tabby 4-payment format) consistently improves CTR for GCC ecommerce categories above AED 200 basket value.
Production specs
- 9:16 vertical, 1080×1920px minimum
- Sound-on hook architecture — audio is the primary hook signal
- Text overlay for sound-off fallback — every key point subtitled
- Hook resolution in first 1–2 seconds
- Creator-native tone — not brand-broadcast tone
- BNPL installment display for GCC ecommerce > AED 200
Platform 02
Meta
Primary metric: ROAS per placement × creative — feed static vs. Reels video measured independently
Meta requires a different production approach per placement. Feed: static images consistently outperform video for direct-response objectives because the thumb-stop moment on a static image is immediate — the viewer sees the offer and the headline in full before deciding to click, rather than needing to watch 3 seconds to understand what is being offered. Static creative on Meta should lead with the offer or the outcome, not the brand. Stories and Reels: vertical 9:16 video with the same hook architecture as TikTok, but with a slightly wider range of acceptable tones — Meta audiences accept a more polished production aesthetic than TikTok without penalising it. Remarketing: dynamic product ads with creative overlaid (price, discount badge, Tabby installment price) consistently outperform generic brand creative for warm audiences who have already viewed the product.
Production specs
- Feed: static 1:1 1080×1080px — offer-prominent headline, minimal copy
- Stories/Reels: 9:16 1080×1920px vertical video
- Remarketing: dynamic product creative with price overlay and discount badge
- Sound-off default — text overlay is mandatory for video
- Separate A/B tests for static vs. video — never assume one outperforms the other
- Arabic-language variant for Arabic publisher and interest targeting
Platform 03
Primary metric: conversion value / cost by asset group — monitored weekly for creative signal
Google creative operates in an intent-driven environment — the creative must match the query intent or audience signal that triggered the placement, not interrupt it. Display remarketing: the creative must reference the specific product or category the user previously viewed — generic brand creative on Google remarketing underperforms targeted dynamic product creative by a significant margin. Performance Max: Google's algorithm assembles the creative from supplied assets (headlines, descriptions, images, video) and optimises toward conversions — asset quality matters more than creative direction at the placement level. The operative creative strategy for Google is: produce a large asset library (10+ headlines, 5+ descriptions, 4+ image formats, 2+ video assets), let the algorithm assemble and test combinations, and monitor which asset combinations the algorithm favours to reverse-engineer the highest-performing creative direction for other channels.
Production specs
- Remarketing: product-specific creative with category reference and offer overlay
- PMax: 10+ headline variants (30 chars max), 5+ description variants (90 chars max)
- Image assets: 1:1 square, 1.91:1 landscape, 4:5 portrait minimum
- Video: 15–30 sec horizontal for YouTube, 9:16 vertical for YouTube Shorts
- Arabic headline and description variants for KSA and UAE Arabic audiences
- Performance Max asset group structured by product category — not by campaign
08 / GCC Localization
GCC ad creatives are engineered for Arabic-language hook architecture, Ramadan creative protocols, and BNPL installment display — not adapted from Western performance playbooks.
Four structural factors make GCC ad creative distinct from global performance creative playbooks: Arabic-language hooks and offer copy that signal local market presence and consistently outperform translated English; Ramadan as a separate creative production track with distinct conversion behaviour and urgency signal requirements; BNPL installment display (Tabby, Tamara) as a conversion lever at the creative level, not only at checkout; and local trust signals (UAE customer counts, GCC-specific payment logos, Arabic reviews) that outperform globally generic equivalents for unfamiliar brands entering GCC markets.
Arabic audience performance
Arabic-language creative — trust, retention, and ROAS
Arabic-language ad creatives consistently outperform English for Arabic-speaking audiences in UAE and KSA — not only in CTR but in hook hold rate and ROAS. The performance delta is not a language preference; it is a trust signal. An Arabic hook communicates local market relevance and brand investment in the market, which reduces the credibility gap for brands that the audience is unfamiliar with. Arabic creative requires native writing, not translation: the hook idioms, offer framing, and CTA copy must be written for the way Arabic-speaking audiences in the Gulf think about value, urgency, and social proof — not translated from English copywriting conventions.
- Arabic hooks outperform English by 30–60% in hook hold rate for Arabic-language audience segments in UAE and KSA
- Hook copy must be written natively — not translated from English headlines
- CTA idioms: 'احصل على العرض' (Get the offer) outperforms translated 'Get Offer' for price-motivated GCC audiences
- Arabic creative requires RTL text overlay placement — not a mirrored version of the English layout
Ramadan and Eid activation
Ramadan creative strategy — seasonal intent and creative protocols
Ramadan produces the highest ad spend volume in GCC markets, but the audience's decision-making framework shifts in ways that require dedicated creative strategy. Gifting intent increases basket size — offer framing around 'the perfect Eid gift' consistently outperforms generic product promotion. Urgency signals around Ramadan and Eid countdowns outperform stock-scarcity signals because the seasonal endpoint is shared, culturally understood, and emotionally relevant. Brand creative associating the product with generosity, community, and family connects with the cultural moment in a way that direct-response product advertising cannot. Ramadan creative is a separate production track — not modified versions of evergreen creative.
- Gifting framing: 'For the people who matter' outperforms product-feature creative for Ramadan activation campaigns
- Eid countdown urgency: culturally understood and emotionally resonant for GCC audiences
- Ramadan creative produced as a separate track — not seasonal overlays on evergreen assets
- Post-Ramadan Eid period: the highest AOV window of the year in most GCC ecommerce categories
Installment display in creative
BNPL in ad creative — Tabby and Tamara as a conversion lever
Displaying Tabby or Tamara installment pricing ('4 payments of AED X') directly in the ad creative — not only at checkout — consistently improves CTR and ROAS for higher-basket ecommerce categories in UAE and KSA. The installment display removes the price objection at the creative level, before the audience reaches the landing page. For baskets above AED 200, the installment framing ('pay AED 75 today') presents a more accessible commitment than the full price ('pay AED 300'). The format that performs best: the Tabby logo + 'starting from AED X' in the offer-prominent section of the static creative or as a text overlay at the conversion layer of the video creative.
- BNPL display in creative increases CTR for baskets AED 200–600 in UAE and KSA ecommerce categories
- Tabby: primary for UAE; Tamara: primary for KSA — produce market-specific creative variants, not one GCC asset
- Installment price in hook: 'Pay AED 75/month for [product]' — removes commitment barrier before landing page
- BNPL creative produces highest uplift for considered purchases (electronics, furniture, fashion above AED 300)
Trust localisation
GCC social proof — local signals outperform global benchmarks
Social proof that works for Western audiences requires localisation for GCC markets. A global customer count ('5 million customers worldwide') provides less trust signal than a UAE-specific count ('50,000 UAE customers') for a UAE audience evaluating an unfamiliar brand. Celebrity endorsements carry higher trust weight if the celebrity is recognisable to UAE and KSA audiences specifically — not global influencers with no GCC presence. Review and testimonial social proof performs best when the reviewer is clearly GCC-based: an Arabic name, a UAE or KSA location, or a review written in Arabic. Local specificity signals market presence and investment, which reduces the perceived risk of purchasing from an unfamiliar brand.',
- UAE customer count outperforms global count for UAE audience trust — by a measurable ROAS margin
- Arabic-language testimonials outperform translated English reviews for Arabic-language audience segments
- UAE and KSA-based influencer endorsements outperform global celebrity creative for GCC performance campaigns
- Local payment logos (Tabby, Tamara, Mada) in creative reduce checkout hesitation for first-time buyers
09 / Creative Programmes We Build
Ecommerce, SaaS, lead generation, and multi-platform. One creative framework.
The ad creative framework is consistent across business models — creative audit, angle hypothesis development, platform-native production, structured test cycles, and quarterly creative iteration. What changes per model: the conversion event (purchase, trial activation, qualified lead), the primary angle hierarchy per audience temperature, the platform weighting, and the performance metric the creative programme is optimised toward.
Ecommerce
Ecommerce ad creative programme
Objective: Purchase ROAS and CPA across TikTok, Meta, and Google for cold and warm audiences
Performance creative for ecommerce — separate angle strategies per audience temperature (cold: problem-first; warm: proof-first; retargeting: offer-first), platform-native formats per channel (TikTok vertical, Meta static + Reels, Google PMax), and a quarterly test cycle that builds four creative hypotheses per cycle and scales winners into the permanent asset library. BNPL display for UAE and KSA categories above AED 200. Ramadan activation as a separate creative track. Arabic-language variants for Arabic audience segments. All tests measured against ROAS and CPA per creative per channel.
Primary metric: ROAS and CPA per creative per channel — weekly
SaaS
SaaS ad creative programme
Objective: Trial activation rate and MQL volume from paid social and search creative
Performance creative for software and subscription — where cold audiences require problem-first hooks ('your paid media attribution is broken — here's why') and warm retargeting requires social proof-first hooks ('how [industry type] companies are using [product] to [outcome]'). Trial-first CTA architecture reduces commitment friction for conversion-stage audiences. Demo CTA for enterprise segments where trial isn't the acquisition pathway. Feature-comparison creative for warm audiences evaluating competing tools. All creative tested against trial_start or demo_request events — not CTR, which doesn't measure acquisition quality for SaaS.
Primary metric: trial activation rate and MQL rate per creative — monthly
Lead Generation
Lead generation ad creative programme
Objective: Cost per qualified lead across finance, real estate, healthcare, and education verticals
Performance creative for lead generation — where qualified lead quality matters more than lead volume, and the creative's job is to filter as much as qualify. Problem-first creative for regulated categories (finance, healthcare, real estate) that establish credibility through mechanism before asking for a form submission. Trust signal integration in the creative: regulatory logos, licence numbers, and data privacy statements visible in the conversion layer for categories where trust is the primary qualification barrier. Arabic-language lead form creative tested independently for Arabic-language audience segments. All tested against cost per qualified lead — not cost per lead, which doesn't measure whether the creative is attracting the right audience.',
Primary metric: cost per qualified lead by creative — monthly
Multi-Platform
Multi-platform creative programme
Objective: Unified creative strategy across 3+ platforms with platform-native production per channel
A creative programme running across TikTok, Meta, Google, and additional channels simultaneously — with a unified creative strategy (consistent angle hypotheses and test cycles) and platform-native production per channel. Creative brief architecture is developed once per angle; platform-specific briefs are derived from the primary brief with channel-specific hook timing, format specifications, and tone guidelines. Test cycles are synchronised across channels so the winner analysis from TikTok informs the Meta test plan and vice versa. Creative library management: a structured asset database tracking which creative is live, which is in test, which is retired, and which is queued — across all channels simultaneously.',
Primary metric: blended ROAS per angle hypothesis — cross-channel — quarterly
10 / Results
One standard: did hypothesis-led creative testing produce measurable ROAS improvement — or did the programme scale untested assets and attribute the outcome to budget rather than creative?
Measured against ROAS improvement and CPA reduction attributable to hook architecture and angle testing — not to changes in budget, audience targeting, or bid strategy. Three ad creative engagements — UAE fashion ecommerce, UAE SaaS, KSA electronics retail — each judged on whether replacing untested, platform-generic creative with hypothesis-led, platform-native creative produced measurably better acquisition economics.
- Fashion EcommerceUAE+186%
ROAS after hook architecture overhaul replaced generic product showcase creative with problem-first hook variants tested per platform — TikTok, Meta feed, and Meta stories independently
A UAE fashion ecommerce brand running the same 30-second brand video across TikTok, Meta feed, and Meta stories — no platform-native production, no hook testing, and ROAS measured at the account level rather than per creative. A creative audit revealed a 3-second view rate of 12% (platform average: 28%) and zero hook architecture documentation. Rebuilding with platform-native vertical video for TikTok, problem-first hooks tested across five variants, and separate Meta static creative for the feed produced a 186% ROAS improvement over 12 weeks — and identified that the hook angle 'why your wardrobe costs more than it should' outperformed the product showcase by 3.4× on TikTok specifically.
cost per acquisition after creative testing identified that problem-first hooks outperformed price-led hooks by 3.4× on TikTok and 2.1× on Meta for new customer acquisition−31%Read the case study - SaaSUAE+94%
trial activation rate from paid social after angle testing identified that social proof-first hooks outperformed problem-first hooks by 2.7× for warm Meta retargeting audiences
A UAE SaaS operator running problem-first creative across all audience segments — cold, warm, and retargeting — using the same hook angle regardless of where in the funnel the audience was. Creative audit showed warm retargeting audiences (seen the brand 3+ times) produced lower trial activation from problem-first creative than from social proof-first creative — because warm audiences had already processed the problem and needed social proof, not problem reiteration. Separating the creative strategy by audience temperature (cold: problem-first; warm: proof-first; retargeting: outcome-first) increased trial activation from paid social by 94% and trial-to-paid CVR by 38% over 10 weeks.
trial-to-paid CVR for paid social cohort after creative angle alignment with the landing page trust sequence produced consistent narrative from ad hook to trial CTA+38%Read the case study - Electronics RetailKSA+61%
hook hold rate after platform-native vertical video with Arabic-language hook copy replaced one-size-fits-all horizontal brand video for KSA TikTok and Meta campaigns
A KSA electronics retailer repurposing horizontal brand video produced for television into vertical TikTok and Meta ad creative — letterboxed, Arabic subtitles added over English audio, and a 30-second format where the hook resolved at 8 seconds into a product demonstration that should have been the opening. Creative production of native vertical video with an Arabic hook ('قبل ما تشتري أي جهاز — اقرأ هذا') and a 3-second product reveal hook, tested against the repurposed horizontal, produced a 61% hook hold rate improvement and a 43% ROAS improvement over the same budget in 8 weeks.
ROAS after improved hook retention qualified higher-intent audiences through to the offer layer — fewer exits at the hook stage, more conversions at the checkout stage+43%Read the case study
Results are reconstructed from server-side tracking and verified attribution. Figures are representative of typical engagements, not guarantees.
11 / Questions
What operators ask about performance ad creative systems before engaging
Questions from ecommerce operators, SaaS businesses, and lead generation brands evaluating a performance ad creative engagement.
Performance ad creative is built for a single measurable outcome: driving a click, a trial start, a purchase, or a form submission. Every creative decision — hook angle, retention structure, offer framing, CTA copy — is made in service of that conversion event and measured against it. Brand creative is built for perception: awareness, association, and recall. The two approaches are not mutually exclusive — performance creative can build brand equity over time — but they start from different hypotheses and are evaluated by different metrics. Performance creative is hypothesis-led (we believe this hook angle will outperform the control by X% because of Y audience psychology); brand creative is judgement-led (this is the correct visual expression of the brand). Performance ad creative requires a test structure, a measurement event, and a significance threshold before a winner is called. Brand creative requires brand guidelines and stakeholder approval. The Adzyon ad creative system operates under the performance model — every creative decision is testable, and every winner is documented before being scaled.
A high-performing hook resolves the scroll-stop problem and establishes relevance in the first 1–3 seconds — before the viewer's pattern recognition decides to skip. Three elements determine hook performance. First, the interruption signal: the first frame must be visually distinct from surrounding content — not a logo, not a product on a white background, and not a title card. Motion, a face, or a text overlay that addresses the viewer directly produces higher 3-second view rates than brand-led openings. Second, the relevance trigger: the hook must signal that the next 15–30 seconds are relevant to this specific viewer's situation. 'Most businesses waste 40% of their ad budget on...' performs better than 'Introducing [product name]' for cold audiences because the relevance is established immediately. Third, the curiosity gap: the hook must create a question in the viewer's mind that the next section of the creative answers — the resolution of that question is the retention mechanism. Hooks that answer their own question in the first 3 seconds produce high view rates and low hold rates — the viewer got what they came for and left. The goal is a hook that creates a question and promises the answer in the next section.
Structured creative testing requires five elements. First, a documented hypothesis: 'We believe a problem-first hook will outperform a price-led hook by 20%+ for cold audiences because the audience is not yet price-motivated at the awareness stage.' Without a hypothesis, the test result is a data point without a learning. Second, a single variable per test: only the hook angle changes between control and variant — format, offer, and CTA remain constant. Testing multiple variables simultaneously produces a result you cannot interpret. Third, a defined primary metric per test stage: 3-second view rate at the hook stage; hook hold rate at the retention stage; ROAS or CPA at the conversion stage. Using the wrong metric at the wrong test stage produces misleading winners. Fourth, statistical significance: 95% confidence level before a winner is called — not 'the variant is trending up' or 'we've spent enough budget to decide.' Fifth, the winner as the new control: the winning variant replaces the control in the next test cycle, so each cycle is testing against the current best performer rather than against the original baseline. This compounding structure is what separates a creative testing programme from a series of one-off experiments.
The formats that perform best are the ones native to each platform's content environment — which is different on TikTok than on Meta. TikTok: 9:16 vertical video, 15–30 seconds, with a hook that resolves in the first 2–3 seconds using a combination of visual movement, direct address, and a text overlay that works for sound-off viewing. The tone must match the creator content around it — performative, direct, and fast-paced. Long production timelines and high production values often produce lower TikTok performance than lo-fi creator-style content shot on a phone. Meta: the highest-performing format depends on placement. Feed placements: static images consistently outperform video for direct-response objectives because they stop the scroll and display the offer in full without requiring the viewer to watch. Story and Reels: vertical 9:16 video with a 3-second hook and offer visible before the skip threshold. Meta does not require creator-native tone — it supports a wider range of creative aesthetics. Google Display: static image creative with offer-prominent copy matched to the audience intent signal — remarketing creative should reference the specific product or category the user viewed, not a generic brand creative.
The correct number depends on the traffic volume available per variant and the metric being tested. The practical rule: enough variants to reach statistical significance in a defined test window, but not so many that you're splitting traffic below a minimum volume threshold per variant. For TikTok and Meta: 3–5 variants per hypothesis is the operating standard. Each variant tests one variable against the control. More than 5 variants per test cycle fragments the traffic budget below significance thresholds for lower-volume campaigns and requires longer test windows to compensate. For lower-budget campaigns (below AED 20,000/month per channel): 2–3 variants per test cycle with sequential testing rather than simultaneous — run variant A against control, reach significance, retire the loser, then run variant B against the winner. For higher-budget campaigns: 4–5 simultaneous variants per test cycle, with spend weighted toward the control until a variant demonstrates clear directional performance, then shifted toward the variant to accelerate significance.
Hook hold rate is the percentage of viewers who continue watching past the hook — typically defined as the percentage who watch 25–50% of a video creative, depending on the platform and format. CTR measures whether the viewer clicked. Hook hold rate measures whether the creative retained the viewer through the argument layer — the section of the creative that builds the case for the offer before the CTA is presented. CTR without hold rate is a misleading metric because a hook with a very aggressive interruption pattern (loud audio, sudden cut, jarring visual) can produce a high CTR from accidental clicks and curiosity clicks that don't convert at the landing page. A creative with 6% CTR and 12% hold rate is producing low-quality traffic; a creative with 2% CTR and 68% hold rate is producing high-intent traffic that has processed the argument and chosen to click. The correlation between hook hold rate and ROAS is consistently higher than the correlation between CTR and ROAS — because hold rate measures intent qualification, and CTR measures only the click.
Ad creative and landing page conversion are the same funnel — not two separate systems. The creative's hook angle creates an expectation in the viewer's mind. The landing page must fulfil that expectation, or the CVR at the landing page is penalised by the creative's own success. A creative with a hook 'we cut our customers' ad costs by 40%' produces an audience that arrives at the landing page expecting proof: case studies, client results, methodology. If the landing page leads with the brand's founding story or a product feature list, the audience's expectation is unmet and they exit. The creative brief and the landing page brief must be developed together — the landing page is the resolution of the creative's promise. The intent type the creative targeted (problem-aware, solution-aware, offer-motivated) determines the landing page content architecture, trust signal sequence, and CTA framing. Creative-to-landing page alignment is a CRO lever as significant as any on-page optimisation.
Three GCC-specific factors affect ad creative strategy. First, Arabic-language creative performance: Arabic-language ad creatives consistently outperform English for Arabic-speaking audiences in UAE and KSA — not only in CTR but in hook hold rate and ROAS. The performance delta is not just a language preference; it is a trust signal. An Arabic hook communicates local market relevance, which reduces the credibility gap for unfamiliar brands. Arabic creative requires more than translation — the hook copy, the offer framing, and the CTA idioms must be written natively, not translated from English. Second, Ramadan creative strategy: Ramadan produces the highest ad spend volume in GCC markets, but the audience's decision-making framework shifts. Gifting intent increases basket size. Urgency signals around Ramadan and Eid timing outperform stock-scarcity signals. Brand association with generosity and community outperforms direct-response framing for brand campaigns during Ramadan. Third, BNPL integration: Tabby and Tamara installment display in ad creative — showing the per-instalment price ('4 payments of AED X') — consistently improves CTR and ROAS for higher-basket categories in UAE and KSA, because the installment framing removes the price objection at the creative level rather than at checkout.
Start with a creative audit
Know which creatives are converting — and which are spending budget behind an untested hook.
A creative audit reviews your current ad library's hook performance per creative, 3-second view rate per angle, hold rate distribution, and ROAS per format — then returns a hook performance heatmap and a first-cycle test plan within five business days. Specific findings: where hook abandonment is costing ROAS before the argument layer is reached, where platform-generic assets are penalised by placement quality scores, and what to test first. No pitch. No commitment beyond the audit.
- Senior creative strategist on every engagement
- UAE · KSA · Global
- Creative audit delivered within five business days