Meta Ads UGC Campaign

A universal Meta Ads funnel — research → ideate → script → cast → produce → compose → launch → measure — with the ppl.studio implementation for each stage. Every step explains what needs to happen, why it matters for the funnel, and how to execute it with our free tools and dashboard features.

Before you start

Have these ready so you don't hit a blocker mid-workflow.

• A product with a public URL (Shopify, Amazon, or your own site).
• A Meta Ads Manager account with the pixel installed (if you plan to launch in Step 12).
• Your product's average order value (AOV), gross margin %, and rough CPM / CTR / CVR benchmarks.
• A ppl.studio account — free tier includes 10 photos; Creator plan ($1.99/week) covers unlimited photos + Animate.

The funnel

Each step is a funnel stage. The top of each step explains what needs to happen in universal marketing terms. The bottom shows how to do it with ppl.studio — with free tools, dashboard features, examples, and gotchas.

1. Step 01~10 min

   Research winning angles in your niche

   Study competitor ads currently running in your category. Extract the angle patterns — not the exact copy — that you'll adapt for your own brand.

   Why this matters — Every successful Meta campaign starts from proven winners, not blank-page creativity. Guessing the right angle adds 4–6 weeks of wasted spend before the pixel finds signal.

   Outcome →A shortlist of 2 angles your audience has already responded to, ready to turn into hooks.

   How to do it with ppl.studio

   ppl.studio's Winning Ads Finder pulls live ads from Meta Ad Library and clusters them into 5 AI-analyzed angles per brand. 10 minutes replaces what used to be a half-day of manual scrolling.

   1. 1.1

      Open the Winning Ads Finder

      No login required. Open the tool in a new browser tab so you can reference it while working through the next steps.

      • Free tool

        Winning Ads Finder — 5 AI-analyzed angles + live Meta Ad Library links for any brand or niche.
   2. 1.2

      Enter 1–3 competitor brands

      Pick direct competitors in your category — same product type, similar price range. The tool fetches their currently-running ads from Meta Ad Library and clusters the patterns into 5 angles.

      Example

      For a protein-bar brand: "RX Bar, Barebells, Quest Nutrition." For a skincare brand: "CeraVe, Cetaphil, The Ordinary."

      Tip: Don't pick aspirational brands ten sizes up — a Louis Vuitton angle won't translate for a $29 candle. Stay in your tier.

   3. 1.3

      Read the 5 AI-analyzed angles

      Each angle comes with its hook pattern (e.g., problem-agitation, social-proof stack), target audience, and emotional trigger. These are patterns, not copy — you'll write your own copy in Step 3.

      Tip: If two angles feel identical, the tool has found a pattern that really works in your niche. That's your highest-probability winner.

   4. 1.4

      Click through to the live Meta Ad Library ads

      Every angle links to the actual running ad. Open 2–3 ads per angle. Watch the first 3 seconds of each video — that's the hook, and that's where 80% of ad performance is decided.

      Tip: Screenshot the three ads that made you stop scrolling. These are your creative reference for Step 7 (photo prompts) and Step 10 (video).

   5. 1.5

      Pick 2 angles to move forward with

      More than 2 angles = unfocused data. Two is enough to tell you which emotional lever your audience actually responds to.

      Tip: Pick angles from different emotional levers (e.g., one fear-based, one aspiration-based). If both win, your audience is broad. If one dominates, you have your lever.

2. Step 02~5 min

   Extract ad concepts from your product URL

   Translate the winning angles into product-specific concepts. Each concept should include a hook, body copy, a CTA, and a visual direction tied to your actual product.

   Why this matters — Angles are generic patterns; concepts make them yours. Without this translation step you'll have good ideas that don't fit what you sell — and buyers won't recognize the product they're clicking on.

   Outcome →3 ready-to-shoot ad concepts tailored to your product.

   How to do it with ppl.studio

   URL → Ad Concepts does the translation automatically. Paste a product URL, get back three concepts with hooks, body copy, CTAs, and visual direction mapped to the exact product.

   1. 2.1

      Open URL → Ad Concepts

      Load the tool and have your product PDP URL ready — the tool pulls the image, price, and description from the page.

      • Free tool

        URL → Ad Concepts — Three ready-to-shoot ad concepts from a product URL — hooks, body copy, CTAs.
   2. 2.2

      Paste your product URL

      Shopify, Amazon, Etsy, or your own site all work. The tool fetches the product image, name, and description, then returns three concepts.

      Example

      https://yourbrand.com/products/signature-candle

      Tip: If a custom e-comm build blocks scraping, paste product name and description manually. The hooks will still be useful.

   3. 2.3

      Review the three concepts

      Each concept comes with a hook, 2–3 lines of body copy, a CTA, and visual direction ("woman holding product on kitchen counter, morning light"). Pick the one matching the angle you chose in Step 1.

   4. 2.4

      Save the concepts as a reference doc

      Copy into a Google Doc, Notion page, or note. You'll reference them for the script in Step 4 and scene prompts in Steps 7–8.

3. Step 03~5 min

   Generate 25 scroll-stopping hooks

   Write 25 opening hooks — the first 1–2 seconds of each ad — then shortlist 3 from different emotional frames. The hook is where 80% of ad performance is decided.

   Why this matters — A weak hook kills the ad in the first 1.5 seconds regardless of the rest. A strong hook buys attention long enough for the rest of the ad to work.

   Outcome →3 shortlisted hooks from 3 different emotional frames, ready to script.

   How to do it with ppl.studio

   The Hook Generator produces 25 hooks grouped by frame — Problem, Curiosity, Social Proof, POV, Bold Claim. Each frame engages a different part of the brain; spreading across frames gives you the test coverage you need.

   1. 3.1

      Open the Hook Generator

      The tool groups 25 hooks by frame. Each frame engages a different psychological lever.

      • Free tool

        Hook Generator — 25 hooks grouped by frame — Problem, Curiosity, Social Proof, POV, Bold Claim.
   2. 3.2

      Describe your product in one sentence

      Keep it literal. "Vanilla-scented soy candle, 9oz, $29" writes sharper hooks than "Luxury atmosphere for your home."

      Example

      Instead of "eco-friendly yoga mat," write "5mm cork yoga mat with alignment stripes, $79."

   3. 3.3

      Scan all 25 hooks in one pass

      Don't judge yet. Read all 25 fast. Note which ones made you pause — those are candidates. A hook that works on you probably works on your audience.

   4. 3.4

      Shortlist 3 hooks from 3 different frames

      Mix frames — don't pick three Social Proof hooks. Spread across Problem, Curiosity, and POV, for example. This tests which emotional lever wins, not which synonym wins.

      Tip: Read each hook out loud. If it sounds like copywriting, rewrite it in plain speech.

4. Step 04~15 min

   Write scripts for your top 3 hooks

   Turn each hook into a 30- or 60-second script with timestamps, on-screen text, shot list, and B-roll notes. The script is the scaffold that turns a hook into a testable piece.

   Why this matters — An unscripted ad is inconsistent and hard to iterate. A scripted ad is reproducible — swap one line and keep everything else constant for a clean A/B.

   Outcome →3 scripts ready to storyboard, each 30–60 seconds with shot list and on-screen text.

   How to do it with ppl.studio

   The UGC Script Generator takes your hook + product and writes the rest in UGC-native voice — not copywriter voice. Run it once per hook.

   1. 4.1

      Open the UGC Script Generator

      You'll run it three times — once per hook. Keep all three outputs in the reference doc from Step 2.

      • Free tool

        UGC Script Generator — 30s or 60s UGC script with timestamps, narration, on-screen text, shot list, and B-roll.
   2. 4.2

      Paste your first hook + product description

      The tool writes the first line of narration verbatim from your hook, then continues into problem, reveal, proof, and CTA beats.

   3. 4.3

      Pick 30s for fast-decision products, 60s for consideration

      30s is default for consumables (food, beauty, candles) — buyers decide fast. 60s is better for higher-consideration purchases (apparel, electronics, skincare routines) where you need to stack proof.

   4. 4.4

      Review the shot list and on-screen text

      Every line has a suggested shot type (wide, medium, close-up, detail) and optional on-screen text. The shot list is the skeleton for your storyboard in Step 9.

      Tip: If the tool gives you 10 shots for a 30s spot, trim to 5–6. Too many shots feels frantic.

   5. 4.5

      Repeat for your other two hooks

      End of step: 3 full scripts. They don't need to be polished — the AI Expert delivers them in voice in Step 10.

   Output: 3 scripts × 30s or 60s — timestamped, with shot list and on-screen text.

5. Step 05~10 min

   Cast your UGC creator

   Lock in the face and voice that will deliver your creative across every asset in this campaign — and every campaign going forward. One creator, consistent everywhere.

   Why this matters — UGC from a single consistent creator outperforms mixed-creator UGC by 20–40% on retention. Consistency builds brand recognition; rotating faces resets trust with every post.

   Outcome →1 AI Expert with a consistent face, voice, backstory, and default wardrobe — reusable forever.

   How to do it with ppl.studio

   In ppl.studio, your "UGC creator" is an AI Expert — a persistent persona with a locked face and voice. Generate one from a niche prompt in ~10 minutes; reuse it across all future shoots.

   1. 5.1

      Go to /personas in the dashboard

      Click "Create expert" or "Add expert." Generate from a niche prompt, or build from scratch field by field.

      • In-app

        AI Experts — Persistent AI personas — consistent face, voice, backstory, expertise, wardrobe.
   2. 5.2

      Enter your niche + product type

      Be specific — concrete niches produce more believable Experts.

      Example

      Good: "30-something new mom, organic baby-food buyer, values transparency." Vague: "Parents who care about health."

      Tip: The face the AI generates first is usually the best — it's tuned to your niche. Only re-roll if something specific is wrong (age off, wrong skin tone for your audience).

   3. 5.3

      Review the generated dossier

      The Expert comes with face, backstory, catchphrases, wardrobe, and default voice notes. Read it end-to-end. Edit fields directly if anything doesn't match your audience.

   4. 5.4

      Save the Expert

      Give it a memorable name (you'll pick it by name in Workbench). Save to your library.

      Tip: For your first brand, one Expert is enough for the whole catalog. Create additional Experts only when serving genuinely different audiences (B2B vs. parents).

   5. 5.5

      Optional: grab the public profile link

      Every Expert has a shareable /experts/[slug] page. Hand the URL to a collaborator or freelancer — they see the persona without logging in.

      • In-app

        Public Expert Profiles — Shareable profile page per Expert — hand the URL to a collaborator.
6. Step 06~5 min

   Prepare product and reference assets

   Upload your product photos, packaging, and any branded accessories. Everything appearing in the ad needs to look exactly like the real product.

   Why this matters — Creative with an approximate product confuses buyers and lowers click-through. Pixel-accurate product rendering converts 3–5× better than approximations.

   Outcome →Product, packaging, and key accessories uploaded, categorized, and ready to drop into scenes.

   How to do it with ppl.studio

   Drop assets into ppl.studio's Props Library once. They'll appear accurately in every future scene — correct colors, exact label placement, right proportions — without re-describing.

   1. 6.1

      Go to /props

      Click "Add props" or "Upload." Upload multiple images per prop — different angles help the AI render accurately.

      • In-app

        Props Library — Upload products once, reuse across every scene with accurate colors and logo placement.
   2. 6.2

      Upload product + packaging + accessories

      Start with: the bare product (front angle), the product in packaging, and any accessories that might appear in scenes (lids, scoops, applicators).

      Tip: White or transparent backgrounds work best. If you only have lifestyle photos, upload them anyway — the AI extracts the product.

   3. 6.3

      Categorize each prop

      Pick "handheld," "large object," or "ambient." Handheld = held by the Expert (candle, bottle). Large object = furniture-scale (car, yoga mat). Ambient = background (plant, laptop).

   4. 6.4

      Name props clearly with SKU

      You'll pick props by name in Workbench. Use SKU-level names.

      Example

      "9oz Vanilla Soy Candle" beats "candle." "Blue Logo Hoodie — Medium" beats "hoodie."

7. Step 07~5 min

   Translate concepts into shoot prompts

   For each ad concept, write detailed shoot prompts specifying scene, pose, lighting, and composition. This is the brief you'd hand to a photographer.

   Why this matters — Vague prompts produce inconsistent photos. Specific prompts produce repeatable, on-brand imagery — critical for ad testing at scale.

   Outcome →6–9 copy-paste-ready prompts (3 concepts × 2–3 scene variants each).

   How to do it with ppl.studio

   The Photo Prompt Generator writes prompts tuned for image models in UGC-specific vocabulary. You plug in the concept and persona; it outputs an optimized prompt.

   1. 7.1

      Open the Photo Prompt Generator

      The tool produces a prompt tuned to the image model — with the right scene, pose, lighting, and composition vocabulary baked in.

      • Free tool

        AI Product Photo Prompt Generator — Optimized Gemini / Nano Banana image prompts. Copy-paste ready.
   2. 7.2

      Pick scene category + persona for each concept

      Categories: "kitchen counter," "gym," "outdoor cafe." Persona: match the Expert from Step 5. The tool blends both into an optimized prompt.

   3. 7.3

      Generate 2–3 prompt variants per concept

      Vary the scene (morning vs. evening light, kitchen vs. bathroom, close-up vs. wide). More variance = more testing signal in Meta Ads Manager later.

   4. 7.4

      Paste all prompts into your reference doc

      Number them (Concept 1 — Scene A, Concept 1 — Scene B, etc.). You'll paste them into Workbench in Step 8 and track winners in Step 12.

8. Step 08~30 min

   Produce lifestyle photos at volume

   Shoot the imagery. Aim for 10–15 photos per concept — volume is the whole game on Meta, where you need variety to find the winning variant.

   Why this matters — Meta's algorithm needs variation to find winners. One-off shots starve the algorithm's learning and leave performance on the table.

   Outcome →30–45 photo variants across your 3 concepts, saved to Gallery.

   How to do it with ppl.studio

   Workbench is the production engine. With your Expert and Props already locked in, each prompt becomes a ~60-second render. Use Visual Presets to keep lighting consistent across the whole campaign.

   1. 8.1

      Open Workbench at /ugc

      Pick your Expert from the dropdown, then open the Props panel and add the product.

      • In-app

        Workbench — Combine Expert + Props. Presets or custom prompts, pose + lighting control.
   2. 8.2

      Paste your first prompt

      The prompt box sits at the top. Paste a prompt from Step 7 — the Expert and Props are locked in, so the AI composes them into the scene.

   3. 8.3

      Tune pose, lighting, location, and shot type

      Workbench has preset controls on each axis. Start with structured presets before free-forming — they're tuned to produce ad-ready output.

   4. 8.4

      Generate 3–5 variants per prompt

      Click generate, wait ~60 seconds, repeat. Every render lands in Gallery automatically. Small differences in pose and lighting give you more ad test cells.

      Tip: Save the best setting as a Visual Preset — one click reuses the same lighting and composition across all future prompts in the campaign.

      • In-app

        Visual Presets — Reference-image presets — save lighting and composition, reuse across shoots.
   5. 8.5

      Move to your next prompt and repeat

      Work through all 6–9 prompts from Step 7. End of step: 30–45 images in Gallery, all tagged to your Expert and product.

   Output: 30–45 photos (3 concepts × 2–3 scenes × 3–5 variants).

9. Step 09~15 min

   Compose carousels

   Arrange photos into 3–5 frame sequences that tell a story and land on a CTA. Meta's carousel placement lets you test multiple hooks within a single ad.

   Why this matters — Carousels have higher engagement than single images in Feed — and the pixel learns from swipes, not just clicks. That's more signal for your budget.

   Outcome →1 carousel storyboard (3–5 frames) per ad concept — 3 storyboards total.

   How to do it with ppl.studio

   Storyboards in ppl.studio lets you either drag-and-drop Gallery frames manually or describe the sequence and let AI generate the full storyboard from one prompt.

   1. 9.1

      Open /storyboards and create a new storyboard

      Two paths: drag-and-drop frames manually from Gallery, or describe the sequence and let AI generate the storyboard from one prompt.

      • In-app

        Storyboards — Multi-frame sequences for carousels and content series.
   2. 9.2

      Pick your hero frame

      Lock the hero (the first frame in the carousel). This is your hook visual — the image that has to survive the scroll test.

      Tip: Match the hero to the hook you wrote in Step 3. If the hook is problem-framed ("tired of dry skin?"), the hero should show the problem.

   3. 9.3

      Fill in the supporting frames

      2–4 additional frames progressing: problem → reveal → proof → CTA. Each lives on its own in Gallery, so you can swap them freely.

   4. 9.4

      Repeat for each concept

      End of step: 3 storyboards total, one per ad concept.

10. Step 10~30 min

    Produce short-form video

    Turn your highest-variance photos into 20–40 second videos with voice and burned-in captions. Meta increasingly favors video across every placement.

    Why this matters — Video has higher reach in Meta's algorithm — especially Reels and Stories. Static-only campaigns leave cheap impressions on the table and can't compete in video-first placements.

    Outcome →3–5 short-form video ad variants (9:16, 20–40 seconds, captions burned in).

    How to do it with ppl.studio

    Animate (VEO 3.1) turns a photo + script phrase into a talking-head video in under 2 minutes. Select your highest-variance photos so video budget is spent on variance, not duplicates.

    1. 10.1

       Go to /animate

       Select a photo from Gallery. Animate is on the Creator plan.

       • In-app

         Animate (VEO 3.1) — Photo → short-form talking-head video with captions. Creator plan.
    2. 10.2

       Pick your highest-variance photos

       Don't animate every photo. Pick the 3–5 frames that differ most (different hooks, different scenes). Video is more expensive to generate; spend budget on variance.

    3. 10.3

       Paste the script's hook line

       Animate speaks the line you paste. Use the first line of each script — that's your hook, and it plays in the first 1.5 seconds of the ad.

    4. 10.4

       Enable burned-in captions

       75% of short-form video is watched sound-off. Captions are non-negotiable. Enable before render.

    5. 10.5

       Render and review

       ~60 seconds per render. Review in Gallery. If voice or mouth feels off, re-render with slightly different script phrasing.

    Output: 3–5 videos, 9:16, MP4, 20–40 seconds each.

11. Step 11~5 min

    Export at platform specs

    Download final creative at every aspect ratio Meta uses: 9:16 for Reels/Stories, 1:1 for Feed square, 16:9 for older placements. Name files for tracking.

    Why this matters — Meta serves the ratio that matches the placement. Single-ratio uploads limit reach across placements. File naming is what lets you identify the winner two weeks from now.

    Outcome →All variants downloaded in 3 aspect ratios, named for performance tracking.

    How to do it with ppl.studio

    Gallery supports bulk export across all three ratios with a naming template — filter by Expert and campaign, then grab everything at once.

    1. 11.1

       Go to /gallery

       Filter by Expert or date to find this campaign's assets quickly.

       • In-app

         Gallery — Every generation lands here. Filter, bulk download, send to campaigns.
    2. 11.2

       Bulk-select by concept

       Select all photos + video for one concept. Meta's algorithm tests at the concept level; you'll group them this way in Ads Manager too.

    3. 11.3

       Download in all 3 ratios

       Export dropdown — 9:16, 1:1, 16:9. Save into separate folders by ratio. Reels + Stories = 9:16. Feed = 1:1 or 16:9.

    4. 11.4

       Rename with campaign + concept + hook convention

       Name like `campaign_concept1_hookA_9x16.mp4`. This is what you'll read in Ads Manager when one variant starts winning.

       Example

       candle-launch_concept2_problem-hook_9x16.mp4

    Output: 9:16, 1:1, 16:9 — MP4 for video, JPG for photos.

12. Step 12~15 min

    Launch, track, and iterate

    Upload variants to Meta Ads Manager, set break-even ROAS as the kill threshold, and schedule the refresh date before creative fatigue hits.

    Why this matters — Launching without a kill threshold burns budget on unprofitable variants. Launching without a refresh schedule lets fatigue silently drive up CPA until you notice weeks later.

    Outcome →Live ad test with break-even ROAS target set and a creative-fatigue refresh schedule in place.

    How to do it with ppl.studio

    Use the ROAS and Creative Fatigue Calculators to set targets before spend goes live, then launch in Meta Ads Manager. The calculators turn business math into specific dates and numbers.

    1. 12.1

       Calculate your break-even ROAS

       Before spending anything, plug AOV, margin, CPM, CTR, and CVR into the ROAS calculator. Output tells you the minimum ROAS you need to not lose money — your kill threshold for underperforming variants.

       • Free tool

         Break-Even ROAS Calculator — AOV, margin, CPM, CTR, CVR → break-even + target ROAS and creative volume needed.
    2. 12.2

       Set up the campaign in Meta Ads Manager

       Create a new campaign. Objective: Sales (or Conversions on older campaigns). Pick CBO for simple tests, ABO for more control.

       • External

         Meta Ads Manager — Launch, pixel, measure.
    3. 12.3

       Upload variants as separate ads within one ad set

       Put all 3 concepts × variants into a single ad set. Let Meta's algorithm find the winner. Don't fragment into multiple ad sets — that starves the conversion signal.

       Tip: Meta needs ~50 conversions per ad set to exit learning. If your budget won't hit that in 7 days, either raise budget or cut variants.

    4. 12.4

       Pick the correct aspect ratio per placement

       9:16 for Reels/Stories, 1:1 for Feed. Meta auto-uses whichever matches the placement — uploading both is fine.

    5. 12.5

       Set a fatigue-refresh schedule

       Run the Fatigue Calculator with expected spend and audience size. Output tells you when frequency will drive CPA up. Block calendar time for that date to refresh creative.

       • Free tool

         Creative Fatigue Calculator — Fatigue score, days until CPA drift, and a creative-refresh schedule.
    6. 12.6

       Launch and monitor

       Launch. Watch at variant level for the first 48 hours. Kill anything 50% under break-even ROAS after 1000+ impressions.

Compare against competitors

These competitors target the same outcome as this workflow — see where ppl.studio wins and where the gaps are.

FAQ

Common questions from first-time runners of this workflow.

Other workflows

Different goal, different funnel.