How to Find Winning Meta Ad Creative: A Signal-Reading Workflow

TL;DR: Finding winning Meta ad creative is a signal-reading problem, not a guessing problem. Run duration is your primary proxy — ads that have been running 30-90+ days are almost certainly profitable. This guide gives you a 4-phase workflow: signal collection, AI-powered pattern extraction, hypothesis clustering, and brief translation. Run it once a month before your next creative sprint.

Every Meta advertiser eventually hits the same wall. You launch five creatives. Two die fast. One limps. Two perform reasonably. You scale the two that perform — but you don't know why they worked, so your next five creatives are a fresh gamble. The creative testing cycle never compounds.

The practitioners who break that cycle do one thing differently: they read the market before they build. They find winning Meta ad creative patterns that already have market proof, then test against those patterns rather than against pure intuition.

This is not surveillance or copying. It's the same competitive research discipline that any product team runs before building a feature. You're reading what's already working in your category so your hypotheses start from a validated baseline rather than zero.

Here's the complete workflow.

What "Winning" Actually Means in the Ad Library Context

Before you can find winning ad creative, you need a working definition of "winning" that doesn't require access to someone else's Ads Manager.

You don't have their ROAS. You don't have their conversion rate. You can't see their CPA. What you have is the single most reliable behavioral signal available in public data: they kept paying for it.

Run duration is a spending decision. Advertisers pause ads that lose money and keep running ads that return profitable results. An ad that has been active for 45 days is not a happy accident — that's a business decision, repeated daily, to continue budget allocation behind that creative.

This doesn't tell you the exact ROAS, but it tells you something more useful for creative research: this format, this hook, this offer structure, this message — this combination works well enough to keep buying.

That's your signal. Everything else in this workflow builds on it.

A note on what run duration doesn't tell you: It doesn't confirm the ad is currently profitable — an advertiser might have paused a 60-day campaign yesterday for strategic reasons unrelated to performance. And it doesn't control for brand-specific factors — a $50M/year DTC brand can afford to run break-even awareness campaigns for months. So treat longevity as a strong hypothesis, not proof. Use multiple data points from multiple competitors before committing to a hypothesis.

Phase 1: Building Your Signal Collection Set

The first phase is deliberate scoping. You're not browsing for inspiration — you're building a structured research set.

Step 1: Define your competitive set.

Select 3-5 competitors who serve an audience segment that overlaps with yours. They don't need to be your direct head-to-head competition. In creative research, even adjacent-category brands can reveal transferable patterns — a DTC skincare brand and a DTC supplement brand often use the same hook structures because they're targeting similar purchase psychology.

For most advertisers, the clearest signal set comes from: 2 direct competitors (same product category, similar price point) + 1-2 indirect competitors (different product, same customer problem) + 1 aspirational brand (the category leader you're eventually trying to beat).

Step 2: Filter for active, long-running ads.

In AdLibrary's unified ad search, search by advertiser name or domain. Apply platform filter to Meta (Facebook + Instagram). Set date range to last 90 days. This surfaces all ads the brand has run in the past quarter.

Now sort by run duration descending — longest-running first. You want ads in the 30-day+ bracket. These are the ones worth close analysis. Filter out anything below 14 days: too recent to confirm as a deliberate scaling decision.

For a typical competitor with active Meta spend, expect 3-8 ads in the 30-day+ bracket at any given time. Across 4 competitors, that gives you 12-32 signal ads to work with — enough for meaningful pattern analysis.

Step 3: Save everything before it disappears.

Meta's Ad Library shows ads for approximately 90 days after they stop running. If an ad paused yesterday, it's still visible today but will be gone in three months. Use AdLibrary's saved ads feature to bookmark each ad in your research set before you analyze — you can return to your analysis session later without the risk of finding that a source ad has expired.

Save all 15-20 target ads before you read a single one deeply. This is the kind of discipline that separates a durable research practice from a one-off session.

Step 4: Log the metadata for each ad.

For each saved ad, record: platform (Facebook feed, Instagram feed, Instagram Reels, Stories), media type (video ads, static image, carousel ads), approximate run duration, and advertiser name. A simple spreadsheet is sufficient. The goal is to know what you have before you start reading for patterns.

Phase 2: AI-Powered Pattern Extraction

Raw ads don't reveal patterns — you have to pull the structure out. This is where AI ad enrichment earns its place in the workflow.

What AI enrichment surfaces:

For each ad, enrichment generates a structured breakdown of: hook structure (the first 3 seconds of video or first line of copy), offer type (discount, free trial, guarantee, social proof, product demo, problem-agitation), emotional trigger (fear of missing out, aspiration, frustration relief, curiosity, urgency), target audience signal (who the ad appears to be speaking to based on language and framing), and call-to-action mechanism.

This is not interpretation — it's extraction. The AI is reading what's structurally present in the ad and naming its components. You're not asking it to speculate about what's working; you're asking it to label what's there.

Why this matters for pattern-finding:

Without structured extraction, pattern analysis is cognitive work you have to do manually for every ad. Across 20 ads, that's 20 separate inference passes — exhausting and inconsistent. With enrichment outputs in a shared document, you can scan a column labeled "hook structure" and immediately see that 7 of your 20 long-running ads open with a problem-statement hook rather than a product-showcase hook. That's a pattern. You couldn't see it from the raw ads.

Enrichment in practice:

For each of your 15-20 saved ads, click into the ad detail view and run AI enrichment. Each enrichment costs 1 AdLibrary credit and caches the result for 30 days — so you can revisit the same session without re-spending credits.

Paste each enrichment output into a spreadsheet row alongside the ad metadata you logged in Phase 1. By the time you've enriched all 20 ads, you have a structured data set — not a folder of screenshots.

For creative teams who run this workflow regularly, the AdLibrary Pro plan at €179/mo gives you 300 credits per month — enough for 3-4 full research sessions covering 15-20 ads each, with room for supplementary searches in between.

Phase 3: Clustering Findings into Hypotheses

Now you read across your structured data set, not down through individual ads.

Cluster by hook type first.

Look at the hook structure column. Group the ads into buckets: problem-statement hooks ("Tired of X?"), outcome hooks ("I went from X to Y in 30 days"), curiosity hooks ("This one thing changed how I think about X"), social proof hooks ("Over 10,000 customers switched to X because..."), product-demo hooks (showing the product immediately in the first frame).

For each bucket, note how many long-running ads fall into it. A bucket with 5+ ads from multiple competitors is a validated hook pattern for your category. A bucket with 1 ad is an outlier.

This is not a vote — it's a signal distribution. If problem-statement hooks dominate your competitor research set and product-demo hooks are sparse, that's the market telling you what resonates with the audience you're both targeting.

Cluster by format second.

Separately group by media type and platform placement. How many of the long-running ads are video vs. static? Short-form (under 15 seconds) vs. longer (30-60 seconds)? Feed vs. Reels vs. Stories placement?

Format patterns matter because they're a proxy for what the Meta algorithm is rewarding. If 8 of your 15 long-running competitor ads are 15-second vertical videos, that tells you something about what format is currently generating efficient delivery in your category — independent of creative quality.

Use AdLibrary's media type filters to cross-check: if you filter only for video ads running 30+ days, are you seeing different patterns than the full set? Sometimes the pattern only emerges when you isolate format.

Form your hypotheses.

For each validated pattern cluster (defined as 3+ long-running ads from different advertisers sharing the same feature), write a one-sentence hypothesis:

• "Problem-statement hooks consistently sustain long runs in this category — test this structure against our current product-demo hook default."
• "Short-form vertical video (10-15 seconds) appears to be the dominant delivery-efficient format — consider a ratio shift toward this format in the next sprint."
• "Social proof as primary offer type (customer testimonials, before/after) appears more durable than discount-led hooks in this category — worth testing a proof-first variant."

Target 2-3 hypotheses per session. More than that and you're trying to test everything at once, which is how you end up with a creative testing backlog that never ships.

Avoid over-indexing on any single brand.

If 5 of your long-running ads are from the same competitor, that's interesting brand intelligence — but it's not a category pattern. A pattern only validates when it appears across multiple independent advertisers. The same hook working for Nike and also for three DTC athletic wear brands is a validated category pattern. The same hook working only for Nike might just be brand equity at work.

For deeper guidance on building systematic research before launching, see pre-launch competitor scan: 30-minute checklist and structuring competitor ad research workflow.

Phase 4: Translating Hypotheses into Creative Briefs

A hypothesis that never becomes a brief is just an observation. Phase 4 is where research value gets captured.

Brief structure for each hypothesis:

For each of your 2-3 validated hypotheses, write a brief that includes: the hook formula (specific, not abstract — "Open with a single frustration statement in direct second person: 'You've been doing X wrong'"), the offer type and social proof requirements, the format specifications (aspect ratio, duration if video, placement target), the CTA mechanism, and the reference ads that validate the pattern (with AdLibrary links from your saved set).

This brief is not a creative direction document — it's a test instruction. The creative team or copywriter has enough structure to build a variant without inventing the strategic direction themselves.

What makes a good brief different from a bad one:

A bad brief says: "Try a problem-statement hook."

A good brief says: "Open with a one-sentence problem statement in the exact voice of the customer: 'I've been spending $3k/month on Meta ads and still couldn't figure out why some creatives worked and others didn't.' Frame the product as the answer to that specific frustration within the first 5 seconds. Reference competitor ad [link] for tone."

The difference is concreteness. The bad brief leaves all the interpretation work to the creator. The good brief encodes the learning directly.

For a full workflow from research session to brief in under 60 minutes, see from ad library research to creative brief in 60 minutes — it covers brief templates you can use verbatim.

Reading the Meta Ad Library vs. Third-Party Tools

Meta provides its own free Ad Library at facebook.com/ads/library. It's the canonical source — every ad running on Facebook and Instagram is visible there, by law. For basic research, it works.

The friction shows up when you try to do the workflow above at any real scale:

• No sort-by-run-duration. Meta's library lets you filter by date range but doesn't give you a "days running" signal on each ad. You have to open each ad individually and calculate manually.
• No saved sets. You can't bookmark ads. Screenshots and browser bookmarks are the only option.
• No AI enrichment. The library shows you the ad. It doesn't extract the structure.
• No cross-platform view. Meta's library only shows Meta ads. If you want to compare how a competitor approaches TikTok versus Facebook, you're switching tools entirely.

Third-party tools like AdLibrary add the infrastructure layer: run duration as a sortable field, saved ad sets, AI enrichment, and multi-platform coverage in one interface.

Meta's free library is fine for one-off checks — verifying that a competitor is running ads, finding a specific ad you saw in the wild. The moment you want to run the 4-phase workflow above with any regularity, you need the operational layer that Meta doesn't provide.

The same principle applies to scale: Meta's free API is adequate for basic programmatic queries against their own platform. When you need to query creative data across Meta, TikTok, and YouTube in a single call — or pull enriched creative intelligence into an automated workflow — Meta's API alone stops being sufficient. AdLibrary's Business plan (€329/mo) includes API access for exactly that use case.

How to Use the Ad Timeline Feature for Deeper Signal Reading

Ad timeline analysis gives you the "first seen" and "last seen" dates plus total days running for each ad. This unlocks signal reading beyond simple run duration.

Refresh cadence as a strategy signal:

If a competitor launches new creative every 2 weeks, they're either testing aggressively (healthy signal) or experiencing creative fatigue fast (their creative quality may be weak, just volume-backed). If a competitor has 3 ads running continuously for 60+ days, they've found a stable scaling creative and are riding it hard.

Both are informative. The first competitor tells you the category requires high creative velocity — your testing cadence needs to match. The second tells you there's a format or message that has real staying power in this category — worth understanding deeply.

Seasonal pattern detection:

Filter for a specific month from last year and look at which creatives were running. Then filter for the same month this year. What patterns recur? Seasonal category dynamics show up in creative patterns — certain hooks, offers, and formats get activated around specific calendar periods. If you can see the pattern a month in advance, you can build to it rather than react.

How to spot a brand discovering its "control" creative:

Watch for a brand that runs many short-duration tests followed by a single ad that breaks out and runs for 30-60+ days. That long-running ad is their control creative — the one they've confirmed works and are scaling hard. Reading that ad carefully tells you what their best current thinking on their audience looks like. It's their most refined message, tested into confidence.

For the full methodology on using timeline signals to read competitor strategy, see diagnosing ad fatigue with competitor longevity signals and reading the Meta algorithm through competitor patterns.

Common Mistakes When Searching for Winning Ad Creative

Mistake 1: Analyzing too many ads from one brand.

If you spend 45 minutes deep on one competitor's ad set, you're learning about that brand, not about your category. Breadth across 3-5 advertisers is more valuable than depth on one. Run duration tells you what works in the market; understanding why one specific brand does what they do is brand analysis, not creative intelligence.

Mistake 2: Treating recent ads as validated.

An ad launched 5 days ago has no performance signal attached to it. It might be brilliant or it might be dead by the time you're reading it. Enforce the 14-day minimum cutoff. If a creative hasn't been running for at least 2 weeks, you're looking at a hypothesis, not evidence.

Mistake 3: Copying format without copying the underlying pattern.

You see a competitor running short-form UGC video and it's been running for 60 days. So you produce short-form UGC video. It dies in a week. What went wrong? You copied the format without reading the hook structure inside that format. The format is a container; the hook is the driver. A problem-statement hook in a short-form UGC video works. A product-demo hook in a short-form UGC video often doesn't. Read the inside of the creative, not just the outside.

Mistake 4: Skipping the hypothesis step and going straight to brief.

If you write a brief for every interesting ad you see, you're producing creative direction at the rate of ad collection. That's volume, not intelligence. The hypothesis step is the filter. Only patterns that appear 3+ times across different advertisers should become briefs. Single-ad insights become notes, not briefs.

Mistake 5: Doing the research after creative production starts.

Competitor research done after your creative direction is set is post-hoc rationalization, not research. It confirms what you already decided. The workflow only adds value if research comes first — before you've committed to a hook direction, before you've briefed the creator, before you've chosen the format. Build it into sprint planning as a mandatory first step.

For how to identify and fix the broader patterns behind weak ad performance, see high-performance ad intelligence creative research platforms and building data-driven creative testing hypotheses from competitor ad research.

Validating Your Hypotheses Before You Spend

Research reduces testing cost — it doesn't eliminate it. Every hypothesis still needs to be tested against your specific audience, your product, and your current conversion funnel. What works for a competitor doesn't automatically work for you. What the research gives you is a validated starting point that's substantially better than a random starting point.

A simple validation structure:

For each hypothesis brief, produce 2-3 variants that apply the validated pattern with different surface executions (different talent, different product angle, slightly different hook phrasing). Run all three in a creative testing sprint with equal budget and a clear kill criterion (e.g., if CPA exceeds 1.5x target after 5 days of data, pause).

If at least one variant from a hypothesis brief outperforms your previous control, the hypothesis was valid — you've transferred market knowledge into your own account. If all three variants underperform, revisit the research: did the pattern actually appear in 3+ independent advertisers, or did you over-weight a single brand?

Tracking the hypothesis lineage:

Keep a log of which briefs came from which research session and which competitive patterns they were based on. Six months later, when you're looking at your ad performance data and trying to understand which creative directions compound, you'll want to know whether your winners trace back to validated category patterns or to internal hunches. That lineage is the most valuable output of a systematic research practice.

For the full framework on structuring creative hypotheses from research, see structured creative research ad hypotheses and analyzing high-performing ad creative framework.

Building the Research Habit: Cadence and Tooling

A one-time research session gives you one set of hypotheses. A monthly research practice builds a compounding intelligence advantage.

Recommended cadence:

• Monthly: Full 4-phase research session (45-60 minutes). Review 15-20 ads across 3-5 competitors. Produce 2-3 new hypotheses.
• Weekly: 10-minute quick scan for any new long-running ads from your top 2 competitors. Flag anything that looks like it's breaking out as a scaling creative — those are your priority signals.
• Pre-sprint: Review your saved hypothesis set before briefing creators. Pull any hypothesis that's been validated internally (ran 14+ days above target CPA) into your "confirmed patterns" library.

Tooling requirements:

At minimum, you need: a tool that shows run duration per ad (Meta's free library doesn't; AdLibrary does), a way to save ads across sessions, and a structured note format for enrichment outputs.

For teams running the full workflow, the AdLibrary Starter plan at €29/mo supports up to 50 credits per month — enough for a single monthly research session with enrichment. The Pro plan at €179/mo (300 credits/mo) supports 3-4 research sessions per month, which is appropriate for agencies running multiple client accounts or in-house teams with active creative programs.

Use the CTR Calculator and CPA Calculator alongside your research sessions to model expected performance ranges for the formats your research identifies — it connects the creative intelligence work to media budget decisions.

The Geo and Platform Dimension

Most creative research treats Meta as a single surface. It isn't.

A hook that works in Facebook feed may not work in Instagram Reels. A format that performs in the US may not perform in Germany. A creative that dominates in one vertical may be invisible in another.

Geo filters let you isolate research by market. If you're launching a campaign in France, your relevant competitive set is brands actively running in France — not US competitors whose creative is tuned for a different cultural register. Run your 4-phase workflow with geo filtering active when your campaign has a specific regional target.

Platform filters let you isolate by placement. Run the workflow separately for Facebook feed, Instagram feed, and Instagram Reels if your campaign will span all three. The winning creative patterns can differ significantly by placement — Reels favors native, fast-cut UGC; feed favors slightly more composed, benefit-led creative.

For brands expanding into multi-platform campaigns, the intelligence gets more valuable (and more complex) when you're reading patterns across Meta, TikTok, and YouTube simultaneously. That's where multi-platform ad coverage becomes the right research scope — looking at how competitors adapt creative across platforms, not just how they perform on one.

Frequently Asked Questions

How do you identify a winning Meta ad creative without access to performance data?

Run duration is the most reliable proxy. Advertisers pause ads that lose money and keep running ads that return profitable results. An ad that has been active for 30-90+ days is almost certainly generating positive returns — otherwise the budget would have been redirected. Filter the Meta Ad Library by advertiser, sort by days running, and focus analysis on creatives in the 30-day+ bracket.

What is the Meta Ad Library and how does it help find winning creatives?

The Meta Ad Library is a public database of all ads currently and recently running on Facebook and Instagram. It shows each ad's creative, copy, and approximate run dates. By filtering for specific advertisers or keywords and sorting by run duration, you can identify which creatives an advertiser has chosen to keep spending behind — which is a strong signal of performance.

What creative patterns should you look for when researching winning ads?

Focus on four dimensions: hook structure (the first 3 seconds or first line of copy), offer type (discount, free trial, social proof, problem-solution), format (video vs. static, aspect ratio, length), and emotional trigger (fear, aspiration, curiosity, urgency). When 3+ long-running ads from different competitors share the same hook pattern, you've found a validated formula for your category.

How many ads should you analyze before forming a hypothesis?

Aim for 15-20 ads per research session across 3-5 competitors. Below 10 ads, patterns are anecdotal — one brand's approach doesn't constitute a validated formula. Above 30 ads in a single session, analysis quality tends to drop because you're working through more variety than pattern recognition can absorb in one pass. A focused 15-20 ad set gives you enough data points to cluster 2-3 reliable hypotheses.

How often should you run a winning creative research session?

For most active Meta advertisers, a monthly research session is sufficient. Run it 1-2 weeks before a major creative sprint so your brief is grounded in current market signals. If you operate in a fast-moving category (fashion, app installs, DTC consumables), a bi-weekly cadence catches trend shifts faster. The goal is refreshing your hypothesis set before your current creative fatigue baseline drops below threshold.

What the Research Tells You That Testing Alone Cannot

Creative testing at scale — running 20 variants per sprint and letting the algorithm pick winners — is a legitimate strategy. It's also expensive and slow to compound. At $500/day in ad spend with 20 variants, you need 5-7 days minimum before any variant has meaningful signal. That's $2,500-$3,500 to learn what a 45-minute research session could have told you before you built a single asset.

The economics are not close. A research-first workflow doesn't eliminate testing — you still need to validate against your specific audience and funnel. But it compresses the discovery phase dramatically. Instead of testing 20 random hypotheses to find 2 winners, you're testing 6 research-validated hypotheses and finding 2 winners. Same number of winners; 70% fewer tests.

Research-first is not a conservative approach to creative. It's a capital-efficient approach. The creative teams that spend the most on testing and the least on research don't win — they just get expensive. The ones who invest in reading the market first build the kind of compounding creative intelligence that actually scales.

For data on how creative research cadence affects campaign performance, see valuing creative time strategy research and high-volume creative strategy meta ads.

Using Competitor Signals to Diagnose Your Own Creative Problems

The 4-phase workflow is primarily designed for prospecting — finding new hypotheses before launching new creative. It's equally useful as a diagnostic tool when your existing creatives are underperforming.

When your current creatives stop working:

Before you assume the audience is exhausted or the algorithm has shifted, check what competitors are running. If your top-performing creative from 3 months ago is losing efficiency and you can see the market has shifted toward a different format or hook pattern, that's not ad fatigue — that's market movement. The category has evolved and your creative hasn't kept up.

This distinction matters because the fix is different. Creative fatigue means replacing the same-pattern creative with fresh production. Market movement means replacing the pattern itself. The first is a production problem; the second is a research problem.

When competitors outperform you on similar creatives:

If a competitor is running a problem-statement hook that looks structurally similar to yours but is clearly scaling harder (longer run duration, more variants of the same pattern), look more closely at the specifics: the exact language of the problem statement, the cultural reference, the casting, the visual pacing of the video. The pattern might be right but the execution might be off. That's fixable without a full creative direction change.

For the full framework on reading competitor performance signals to fix your own account issues, see guide to analyzing competitor ad creative strategies and a strategic guide to pruning and refining ad creative.

Scaling the Research Practice Across a Team

For solo operators, the 4-phase workflow runs in under an hour. For teams, it can become a shared intelligence practice that compounds across every campaign and account.

Shared research libraries:

When multiple team members run research sessions, the output should go into a shared asset: a structured database of validated hypotheses with supporting ad evidence, organized by category, hook type, and validation date. New team members onboard into a library of proven patterns rather than starting from zero.

Assigning research ownership:

In an agency or in-house team, assign competitive research ownership by vertical or account. One person runs the monthly research session for their category portfolio, enriches the ads, clusters the hypotheses, and presents 2-3 briefs to the creative team. The creative team's job is execution, not strategic direction discovery.

Connecting research to results tracking:

After each test sprint, log which briefs produced winners and which didn't. After 6 months of this data, you have an internal record of which category patterns predict winners in your specific funnel — a proprietary intelligence layer no competitor can replicate.

For operational patterns on running a high-volume creative research and testing practice, see creative strategist research workflow with an ad library and how to reverse engineer winning ads creative strategist playbook. Also see creative-strategist-workflow for the full workflow model in a single reference.

The Role of External Sources in Validating Creative Patterns

Competitor ad research tells you what's working in your immediate category. It's the strongest signal available — but it's not the only one worth reading.

Two external sources worth cross-referencing:

Meta's own Creative Guidance: Meta's Business Help Center publishes data on format performance benchmarks — video vs. static, recommended aspect ratios, optimal video lengths by placement. This is aggregate data across all Meta advertisers, not your specific category. Use it as a floor check: if your research is pointing toward a format that Meta's own data suggests underperforms, investigate why before committing.

IAB Creative Standards: The Interactive Advertising Bureau's creative standards documentation provides category-specific guidance on ad quality signals that affect delivery auction outcomes. Understanding what makes an ad score well on Meta's relevance diagnostics helps you design creatives that don't just have strong hooks — they also get efficient delivery.

HubSpot's annual State of Marketing report (hubspot.com) tracks format preference and content consumption trends across platforms annually. Cross-referencing your category research against broader format trends can surface whether a pattern you're seeing is specific to your niche or part of a larger market shift.

Combining competitor intelligence (what's specifically working in your category) with platform guidance (what formats deliver efficiently) and industry benchmarks (what trends are driving broader format shifts) gives you a triangulated view. Each source fills gaps the others can't.

Getting Started: Your First Research Session in 45 Minutes

If you haven't run a structured research session before, here's the exact sequence for your first one:

1. Minutes 0-5: Open AdLibrary's unified ad search. Search for your top 3 competitors by name. Apply Meta platform filter and 90-day date range.
2. Minutes 5-15: For each competitor, sort by run duration. Save all ads running 14+ days to your saved ads library. Target 5-7 per competitor, 15-20 total.
3. Minutes 15-30: Run AI enrichment on all 15-20 saved ads. Paste each output into a spreadsheet row with columns: advertiser, run duration, format, hook type, offer type, emotional trigger.
4. Minutes 30-40: Read across your spreadsheet. Circle any hook type, format, or offer structure that appears 3+ times from different advertisers.
5. Minutes 40-45: Write one hypothesis per validated pattern. These are your briefs for the next sprint.

That's it. The first session takes longer because you're setting up the spreadsheet and getting familiar with the workflow. By session three, you're doing it in 35 minutes.

For the Starter plan at €29/mo (50 credits/month), a session of 20 saved ads + 20 enrichments uses exactly 20 credits — leaving 30 credits for supplementary search and analysis throughout the month. The ROAS Calculator can help you model what improvement in creative efficiency justifies the subscription cost in your specific budget context.

The Bottom Line

Finding winning Meta ad creative is not a creative talent problem. It's a signal-reading problem. The market runs experiments daily and makes its findings public — in the form of long-running ads that advertisers keep paying for. The 4-phase workflow gives you a repeatable method for reading those findings and converting them into briefs your team can build against.

The practitioners who build this as a monthly habit consistently outperform those who rely on intuition or isolated testing. Not because they're more creative, but because they start every sprint from a validated baseline rather than from zero.

Start with one session. Pick 3 competitors. Save 15-20 ads. Run enrichment on each. Read for patterns. Write 2 hypotheses. Ship the briefs.

Do that once a month and your creative testing compounds. Skip it and your testing resets to zero every sprint.

The AdLibrary Pro plan at €179/mo gives you 300 credits per month — the right scale for a consistent research practice without rationing. If your workflow scales to multi-platform research or programmatic intelligence queries, the Business plan at €329/mo includes API access for the same intelligence layer in automated pipelines.