Creative Systems Agency · Dubai · UAE · KSA
Creative that tests,
wins, and
compounds.
A performance creative operating system — strategy, production, and testing infrastructure that generates winning ad creative at the velocity algorithmic platforms demand.
3.1×
Median ROAS lift from creative iteration
48 hrs
Avg. hypothesis to live test
200+
Winning variants produced & scaled
02 / The Creative Bottleneck
Creative is the growth lever. Most brands treat it as a cost.
In a world where platforms auto-optimize bids and audience targeting, creative quality is the primary variable operators can still move. The brands that compound on paid media are the ones that test creative at velocity. The ones that stall are running the same ads from three months ago.
3.4 wks
Avg. performing creative lifespan before fatigue
Without a systematic refresh protocol
1 in 8
Creative variants that outperform control on first test
Testing volume is the primary driver of improvement
+81%
CPP increase at creative fatigue vs. refresh threshold
Industry average across ecommerce accounts
These are structural problems — not production problems. Hiring more designers does not solve a testing infrastructure gap. More creative volume without a hypothesis framework produces more noise, not more winners. The constraint is the system, not the output.
Creative fatigue erodes ROAS without a visible signal
Algorithms continue spending at the same pace as creative effectiveness declines with rising frequency. Ad accounts show stable spend while conversion metrics degrade — no alert fires in the platform.
ROAS degrades over weeks and gets attributed to seasonality or competition. The real cause — stale creative — is invisible without frequency-level creative analysis.
Production velocity can't match the testing demand
Algorithmic platforms require rapid creative rotation to maintain performance. Internal teams and standard agencies produce on brief-cycle timelines — weeks, not days. Testing stalls before it starts.
Winners are never identified because tests never complete. Creative decisions default to consensus and aesthetic preference — consistently underperforming the actual ROAS ceiling.
Creative is selected by preference, not performance data
Without a hypothesis-led testing framework, creative variants are evaluated by team consensus, brand guidelines, or aesthetic judgment — none of which correlate consistently with conversion.
The creative that looks best in a deck performs worst in an auction. High-production-cost polished assets underperform lo-fi hook-led variants that nobody would have approved.
03 / The Creative System
We run creative as an operating system, not a production batch
Five stages from creative audit to compounding brief — each producing the output the next requires. Creative Audit surfaces what performance data says is working, independent of what the team thinks is working — the foundation that makes brief development evidence-based rather than instinct-based. Brief Development translates those findings into testable hypotheses before production begins. Multi-Format Production executes those hypotheses to platform spec — volume determined by testing cadence, not batch convention. Controlled Testing evaluates one variable at a time against defined thresholds, with no early calls on insufficient data. Scale & Refresh enters winners into the scaling protocol and updates the brief framework, so the next cycle begins from a stronger baseline than the last.
Why strategy and testing infrastructure must precede production volume
Most creative programmes are built backwards — production first, strategy and testing retrofitted. By the time structural issues are visible in the ROAS data, production has continued in the wrong direction for months. We establish the brief framework and testing infrastructure before committing to production volume.
- 01
Creative Audit
Analyse the full creative library: historical performance by format, hook type, visual treatment, and audience. Identify fatigue state, format coverage gaps, and what the data says is working versus what the team thinks is working. Most libraries have one or two buried winners obscured by campaign structure.
Output: Creative performance audit - 02
Brief Development
Performance briefs built from data — not from brand decks. Winning hooks, proven angle types, format gaps, and audience-creative alignment translated into testable hypotheses. The brief is the specification; production executes to it. Briefs that don't come from data produce guesswork at scale.
Output: Data-led creative brief framework - 03
Multi-Format Production
Static, motion, UGC-style, carousel, short-form video — produced to platform spec across the defined test hypotheses. Production volume is determined by testing cadence, not by what looks complete in a deck. Every asset is produced to a defined hypothesis, not to fill a batch.
Output: Test-ready creative variants - 04
Controlled Testing
Structured creative tests with defined hypotheses, success metrics, and minimum impression thresholds for statistical confidence. One variable per test. Results evaluated only after sufficient data — no early calls on noise. Winners are identified, documented, and explained — not just flagged.
Output: Validated creative winner - 05
Scale & Refresh
Winning creatives enter a scaling protocol across placements, formats, and budgets. A refresh calendar maintains creative freshness before fatigue sets in — not reactively after ROAS degrades. Learnings update the brief framework. The next cycle begins from a stronger creative baseline than the last.
Output: Compounding creative system
Want to see how this applies to your funnel?
A senior strategist reviews your specific setup — complimentary, no pitch deck.
04 / What We Build
Four capabilities. One creative operating system.
Creative strategy defines the hypotheses. Testing infrastructure validates the winners. Performance design and motion creative produce the variants that feed both. The four capabilities operate as a single system — not as separate services that require manual coordination.
The brief that makes production worth doing
Creative Strategy
Creative strategy is the upstream layer that determines whether production time is invested against the right hypotheses. Hook angle development, competitive creative audit, audience-creative alignment, and format prioritisation — done from performance data, not brand intuition.
- Hook angle development and testing
- Competitive creative landscape audit
- Audience-creative alignment mapping
- Format and placement strategy by channel
The system that finds winners systematically
Creative Testing Infrastructure
Testing infrastructure is what separates a performance creative partner from a production house. Hypothesis-led testing cadence, A/B variant structure, winner identification criteria, and a brief framework that updates with every completed cycle — the operational system that makes creative compounding possible.
- Hypothesis-led creative test structure
- A/B variant production and deployment
- Winner identification and documentation
- Creative performance reporting dashboard
Static and display creative built to convert
Performance Design
Direct response format design for Meta, TikTok, Google, and programmatic display — produced to platform spec and optimised for conversion, not brand aesthetics. Product-first visual hierarchy, high-contrast CTAs, and format variants calibrated to placement requirements.
- Direct response static creative (1:1, 4:5, 16:9)
- Carousel and catalogue creative for ecommerce
- Google Display and Performance Max assets
- UGC-style static and text-overlay formats
Short-form video and motion assets for performance
Motion & Video Creative
Short-form video, motion graphics, and UGC-style creative produced for TikTok, Reels, YouTube, and connected TV. Hook-and-reveal formats, problem-led narrative structures, and brand-safe native production — built for algorithmic distribution, not broadcast.
- Short-form video 6s, 15s, 30s (9:16 and 1:1)
- Motion graphics and animated display
- UGC-style narrative and testimonial formats
- Hook testing across first-3-second sequences
The system is deployed as an integrated programme — creative strategy informs the brief, production executes to the brief, testing validates against the hypothesis, and winners enter the scaling protocol. Each cycle makes the next brief smarter.
05 / Testing Framework
Five stages. Every creative test. No shortcuts.
Every creative test runs through the same protocol: observe, hypothesize, produce, test, scale. Speed does not justify skipping the observation or minimum-sample stages — that is how creative decisions get made on noise and cost you ROAS instead of improving it.
Observe
Platform analytics, frequency-level fatigue data, hook drop-off rates, and audience response patterns across the active creative library.
Hypothesize
One testable variable — hook angle, opening frame, visual treatment, format type, or CTA. Defined success metric. Minimum impression threshold set before the test begins.
Produce
Multi-format creative produced to hypothesis spec — platform compliant, in the volume required to run concurrent tests without audience overlap.
Test
Controlled live test against the defined hypothesis. Results evaluated only after minimum impression volume — no early calls on insufficient data. No opinion at this stage.
Scale
Winner enters the scaling protocol across placements and budgets. Learnings update the brief framework. The next cycle begins from a more informed creative baseline.
The brief framework updates with every completed cycle — each test makes the next hypothesis sharper. A creative programme running this framework for 12 months operates from a materially deeper understanding of what converts than one running ad-hoc production.
Distribution layer: Paid Media
06 / Platform Creative
TikTok, Meta, and Google each demand a different creative strategy.
A single creative approach applied across platforms consistently underperforms. TikTok rewards hook velocity. Meta rewards format diversity. Google rewards intent alignment. We produce platform-native creative — not cross-posted repurposing.
TikTok
Native-first. 3-second or nothing.
TikTok's algorithm distributes based on early engagement signals — hook quality determines whether creative reaches meaningful scale. Native-style production, trend-aware visual language, and UGC-format authenticity consistently outperform polished brand creative. The hook is the ad.
Meta
Prospecting and retargeting need different creative.
Meta's creative ecosystem spans prospecting and retargeting with distinct format requirements at each stage. Catalogue and DPA formats serve ecommerce differently from single-image direct response. Each placement demands a distinct creative strategy — there is no universal Meta creative format.
Asset bundles. Responsive assembly. Search intent.
Performance Max requires a complete creative asset bundle — headlines, descriptions, images, and video — assembled responsively across placements. Search intent alignment and responsive creative composition require a different production approach than social-first creative.
Creative tests are run per-platform, not shared across channels. A winning hook on TikTok does not automatically transfer to Meta — the algorithms, audience behaviour, and creative expectations are structurally different. Platform-specific testing produces significantly higher win rates than cross-posted creative.
GCC & MENA
Gulf market expertise, not Western campaigns translated.
UAE and KSA audiences behave differently from Western markets — different platforms dominate, different creative formats convert, and different compliance constraints govern what a funnel can do. We build acquisition systems for the region, not global templates re-skinned with Arabic text.
Most international agencies entering the GCC bring a proven Western playbook and adapt it. The adaptation usually means translating copy and switching the flag on the geo-targeting. It does not address platform hierarchy (Snapchat is a primary channel in KSA, not an afterthought), CTA architecture (WhatsApp outperforms form flows in relationship-first buying cultures), or attribution (UAE and KSA require separate measurement pipelines, not a merged GCC total).
- Arabic-first creative built from brief, not translated from English
- Platform mixes calibrated to Gulf audiences — Snapchat and TikTok outperform Meta in KSA demographics that Western agencies undervalue
- Compliance-aware funnel design for regulated verticals — finance, healthcare, and forex in UAE and KSA
- UAE and KSA attribution maintained as separate pipelines — not merged into a single 'GCC' number that hides market-level differences
- WhatsApp conversion architecture for markets where form-first funnels structurally underperform
GCC and MENA campaigns managed across ecommerce, SaaS, finance, real estate, and healthcare. Arabic and English campaigns share the same attribution system and are reported against a single efficiency target — with UAE and KSA maintained as separate pipelines to surface market-level differences that a blended GCC number would hide.
08 / Creative + Tracking Feedback Loop
Creative that can be measured compounds. Creative that can't, stalls.
Variant-level attribution turns creative testing from an opinion into a system.
Creative-level attribution
Conversion events matched to specific creative variants — not just to campaign or ad set. Without variant-level attribution, you can identify that a campaign works but not which creative is responsible. One piece of information is actionable; the other is not.
Frequency and fatigue signals
Server-side frequency data that identifies creative fatigue at the variant level before ROAS degrades. Knowing a creative has hit frequency 4.8 on your highest-value segment tells you when to refresh — not after the CPP has already increased 40%.
Incrementality by creative variant
Holdout testing that measures genuine uplift of winning creative variants versus control — identifying which variants actually added new customers versus captured existing intent that would have converted regardless.
Measurement infrastructure for creative performance: Tracking & Analytics
09 / Results
One standard: did hypothesis-led creative testing produce compounding acquisition improvement — or did performance reset with each new production batch?
Measured against CTR improvement and CPP reduction attributable to the creative testing system — not to changes in budget, audience targeting, or bid strategy. Three creative systems engagements — UAE ecommerce, KSA finance, global SaaS — each judged on whether winning creative compounded as the testing framework matured, and whether production continued in the direction the data specified.
- EcommerceUAE+180%
CTR uplift from creative iteration
Rebuilt creative brief framework, introduced hook angle testing across three angle types, and produced 24 variants in the first 90 days. The winning hook increased CTR 180% and reduced CPP 34% vs. the static creative it replaced. Programme now runs a 6-week creative refresh cycle.
Cost per purchase vs. control period−34%Read the case study - FinanceKSA+290%
Qualified lead volume from creative testing
Tested five creative angles for a KSA finance brand — fear-led, aspiration-led, comparison-led, social proof, and product-led. Aspiration combined with social proof outperformed all others by 290% on qualified lead volume. All subsequent creative produced from this framework.
Cost per qualified lead−47%Read the case study - SaaSGlobal3.8×
Trial activation: problem-led vs. feature-led creative
Creative strategy test: problem-led video (opening with the user's pain state) vs. feature-led (opening with product capabilities). Problem-led outperformed by 3.8× on trial activation rate. All subsequent creative produced problem-led. Feature-led creative moved to retargeting only.
Video creative CTR on LinkedIn and Meta+220%Read the case study
Results are reconstructed from server-side tracking and verified attribution. Figures are representative of typical engagements, not guarantees.
10 / Questions
What operators ask about performance creative systems before engaging
Straight answers on creative production, testing infrastructure, platform-specific creative strategy, GCC creative localization, and how creative performance connects to tracking and paid media.
A creative systems engagement includes a full creative audit of your existing library, a brief development framework built from performance data, multi-format creative production across your active channels, a structured testing programme with defined hypotheses and success thresholds, winner identification and scaling protocol, and an ongoing creative refresh cadence. You receive a creative performance dashboard showing test results, winning variants, and creative fatigue indicators — updated as each test cycle completes.
Creative testing requires a minimum of two to four weeks to collect statistically reliable data per test — rushing to conclusion on insufficient data is how creative decisions get made on noise rather than signal. If you have an existing creative library, we typically identify performance improvement in the first test cycle, which completes within 30–45 days. If creative is being built from scratch, the first meaningful test results come in at 45–90 days. We will set clear expectations based on your current creative state and platform spend levels.
Creative testing becomes reliably ROI-positive at monthly platform spend above approximately AED 50,000 ($14,000 USD). Below this, impression volumes are too low to reach statistical confidence within a reasonable test window. Above this threshold, the cost of systematic creative testing and production is consistently recovered through ROAS improvement from winning creative. The higher the spend, the faster tests complete and the faster winners compound.
Standard creative agencies produce deliverables — files, batches, assets. We build the creative operating system: a brief framework derived from performance data, a testing cadence that validates what works before scaling it, and a creative refresh protocol that maintains performance over time. The structural difference is that our creative gets more effective as each test cycle completes. Standard production delivers the same quality output indefinitely — ours compounds.
Yes. GCC creative has specific requirements that most global agencies mishandle: Arabic-language creative requires native production — the visual hierarchy, reading direction, and cultural reference points are different from translated assets; Ramadan and seasonal creative windows follow a distinct GCC calendar that affects hook angles and creative timing; BNPL payment messaging (Tabby, Tamara) is a material conversion driver in UAE and KSA and needs integration into creative strategy; and regulated verticals (finance, healthcare) have compliance constraints that affect visual and copy approach. We produce creative natively for GCC markets.
We produce static creative (single image 1:1 and 4:5 for Meta, display banners for Google), carousel and catalogue creative for ecommerce, short-form video (6s, 15s, 30s for TikTok, Reels, YouTube), UGC-style video creative and testimonial formats, motion graphics and animated display, and Performance Max creative asset bundles for Google. Format mix is determined by platform and testing strategy — we do not produce one format type across all channels.
Each creative test is built around a defined hypothesis: a specific variable (hook angle, visual treatment, format, CTA) against a defined success metric (CTR, CVR, CPP, or ROAS). Tests run against a minimum impression threshold before results are evaluated — we do not call winners on low-volume data. Winning creative is identified against the hypothesis metric, documented in the brief framework, and enters a scaling protocol. Underperformers are paused and analysed for learnings that inform the next hypothesis.
Creative performance measurement requires the same infrastructure as any other paid media signal: server-side event tracking, accurate platform Conversions API integration, and first-party attribution. A creative variant that appears to lose in-platform may be winning on incremental new customers when measured against first-party order data. We build creative systems on top of tracking infrastructure — and where that infrastructure is missing, we flag it before we produce. The feedback loop between creative performance data and the brief framework is how creative iteration compounds over time.
Relevant for
Start with a creative review
Find out exactly where your creative is leaving ROAS on the table
Book a 30-minute call and we will audit your current creative library — identify fatigue state, format gaps, winning hook patterns, and untested hypotheses — then deliver a written creative assessment within five business days. Specific findings: where creative fatigue is eroding ROAS without a visible platform signal, where production is running without a hypothesis framework, and what to test first. No pitch. No commitment beyond the review.
- Senior creative strategist on every call
- Written creative audit in five business days
- UAE · KSA · Global markets