Paid Media Build Kit

Launch Budget & Channel Split

Assumption: USD 60/day starting paid budget for the launch week Sep 8–14 (W3 of the cold-start calendar, paid go-live Monday Sep 8). Weekly cap = USD 420. All CPAs below are cost-per-completed-signup (Lead), not cost-per-click; SEA e-commerce seller traffic is mobile-first and cheap on CPM, so targets are aggressive but achievable with Instant Forms and a frictionless free signup.

Spend ramp (weeks 2–4 of paid): - Wk1 Sep 8–14 — $60/day (this plan). Learning phase. No scaling decisions until an ad set exits learning (50 conversions) or 5 days elapse. - Wk2 Sep 15–21 — $100/day. Unlock: blended signup CPA ≤ $3.00 for 3 consecutive days AND first-search completion rate (signup→FirstSupplierSearch) ≥ 45%. Adds a second country to Meta and turns on 11.11 seasonal keywords. - Wk3 Sep 22–28 — $150/day. Unlock: lookalike (LAL 1–3% of signups) beats interest audiences on CPA, AND ≥ 20% of signups come from retargeting at CPA ≤ $1.50. Adds Google Demand Gen + TikTok Prospecting. - Ramp is throttled to $60/day again if blended CPA breaches $4.00 for two straight days.

Meta Ads Account Build (ready to create)

Campaign structure

Naming convention: meta_[country]_[objective]_[audience]_[creative]_[date]. Country = sea for multi-country ad sets. Objective selection: Sales once the Pixel/CAPI Lead event has >100 lifetime events (optimizes to real signups); Lead for Instant Form campaigns (optimizes to form opens even on thin data); Traffic only for the cheapest possible top-funnel warming in new countries for the first 3–4 days before switching to Sales — never run Traffic as the permanent objective.

Country behaviour note: VN, MY, ID skew to mobile-only, low-card-usage, and WhatsApp/Zalo-based buyer conversations. Do not ask for credit card or long forms there; Instant Form or WhatsApp click-to-chat outperforms a landing-page lead form by a wide margin. TH and PH tolerate a light LP with more copy.

Audience list (12+ concrete, build as saved audiences)

Exclusions on every prospecting set: Custom Audience of existing registered users (email/phone list), CompleteRegistration site event 180d, employee/device list, and interest-based exclusion of seller-center job seekers (exclude Job hunting, Resume, Internship, Freelance job interests).

Broad test cell-set recommendation: run one ad set with no interests at all (Audience #10) at ≥ $6/day. Meta's post-2023 delivery is creative-driven; the broad set is the control that tells you whether your interest stacks are actually helping. Keep it live for the full launch week before judging.

Placements & creative specs

Use Advantage+ placements but manually cap to: Facebook Feed, Instagram Feed, Instagram Reels, Instagram Stories, Facebook Reels. Turn off Audience Network and Messenger (junk clicks). - Feed: 1:1 (1080×1080) or 4:5 (1080×1350). Keep text overlay under 20% of frame. - Reels/Stories: 9:16 (1080×1920), 15s max, leave top 250px and bottom 350px free of key text/UI (safe zone for Reels caption bar and Stories sticker zone). Hook must land in 0–3s. - Reels duration sweet spot for this product: 8–12s screen-recording of a search returning suppliers.

3 ad sets × 3 creatives (9 total, English)

Angle key: (a) 9.9 selection anxiety · (b) price/MOQ transparency · (c) scam/fake-supplier fear & verification · (d) time saved vs manual 1688 · (e) competitor/alternative · (f) free-during-9.9.

Ad Set 1 — meta_ID_sales_lal-signups_99selection_20260908

• 1A (angle a, emoji-free) — Primary: "9.9 is in 7 days. If you still don't know what to stock, that's the problem — not your effort. Type a product idea or drop a photo and see live supplier prices, MOQ, and demand signals in seconds." Headline: "What to stock for 9.9, solved" Description: "Free AI sourcing search for Shopee & TikTok sellers."
• 1B (angle f, emoji-light) — Primary: "🔥 Sourcing search is FREE through 9.9. Upload a product photo, get the source factories behind it — price, MOQ, and what's actually selling. No credit card." Headline: "Free until 9.9 — find suppliers fast" Description: "Type an idea. Get real factory prices."
• 1C (angle b, emoji-free) — Primary: "Guessing the wholesale price wastes your 9.9 margin. See the real factory price AND the minimum order quantity for any product, side by side across suppliers." Headline: "Real price + MOQ, no back-and-forth" Description: "Compare source suppliers in one search."

Ad Set 2 — meta_SEA_sales_broad_sellers_99selection_20260908

• 2A (angle c, emoji-light) — Primary: "Half the 'factories' on big sourcing sites are middlemen — or worse. Every supplier we surface is scored on years trading, transaction history, and verification, so you don't get burned before 9.9. ⚠️" Headline: "Find real factories, not resellers" Description: "Verified suppliers + risk signals built in."
• 2B (angle d, emoji-free) — Primary: "Manually scrolling 1688 for one product line takes an afternoon. Our AI does the supplier shortlist in ~20 seconds — price, MOQ, demand. Reclaim your week." Headline: "Stop scrolling. Start sourcing." Description: "AI search over source suppliers."
• 2C (angle e, emoji-light) — Primary: "You've tried the big sourcing sites and still can't compare suppliers fast. This is the search layer on top of them: one question, ranked factories, real numbers. 🛒" Headline: "A faster way to find suppliers" Description: "Built for SEA sellers, not for Chinese buyers."

Ad Set 3 — meta_SEA_lead_interest-sellers_pricemoq_20260908

• 3A (angle a+f, emoji-light) — Primary: "9.9 countdown is on. ⏳ Don't pick products blind. Free AI sourcing search until Sept 9: photo → suppliers → price + MOQ. See what's hot before you spend." Headline: "Free until 9.9: pick winning products" Description: "Type or upload, get supplier truth."
• 3B (angle b+d, emoji-free) — Primary: "The two numbers that decide your margin are price and MOQ. We show both for matching source factories in one search — no translation, no chasing sellers on chat." Headline: "Price + MOQ in one search" Description: "Cut a day of manual sourcing to 20 seconds."
• 3C (angle c+a, emoji-free) — Primary: "Stocking for 9.9 with an unverified supplier is how you lose twice — bad goods, missed sale. Every result is checked for trading history and verification before you see it." Headline: "Verified suppliers before 9.9" Description: "Reduce scam risk on your next order."

Forms / CTA & lead capture

Prefer Instant Form for the VN/MY/ID interest set and all Traffic→warm retargeting: lowest friction, no page-load drop-off, cheapest CPB. Fields: Full name, Phone number (WhatsApp-prefilled), Business type (Shopee/Lazada/TikTok/offline dropdown), One open-text "what are you sourcing?" — that last field pre-seeds the first search. Prefer the landing page for the Sales-optimized LAL and broad sets, because you can fire the custom FirstSupplierSearch event and drive real product activation, not just form opens. Rule of thumb: capture lead cheaply with forms, but push intent-heavy LAL/broad audiences to the LP where a signup converts into a first search. CTA button: Sign Up on Sales, Get Quote/Learn More avoided for cold (use Get Offer on the free-until-9.9 set).

Tracking checklist (do before spend)

• Pixel events to fire: PageView, Lead (form submit), CompleteRegistration (account confirmed), Search (any product query submitted), and custom FirstSupplierSearch (first query that returns ≥1 supplier — this is the real activation metric).
• Conversions API: install server-side CAPI alongside browser Pixel, match on hashed email+phone, deduplicate with event_id. Set test events in Events Manager, confirm browser+server both receive.
• Domain verification in Business Manager (DNS TXT) before configuring CAPI, or aggregated events will stay blocked on iOS.
• Configure Aggregated Event Measurement, priority order: CompleteRegistration > FirstSupplierSearch > Lead > Search > PageView.
• Event match quality target: ≥ 6.0/10 on both Pixel and CAPI before launch; below 6.0 fix identity matching, do not scale.
• UTM template: utm_source=meta&utm_medium=paid&utm_campaign={campaign_name}&utm_content={ad_set}.

Google Ads (finish the build started in W2)

Match types: "phrase" / [exact] shown. All campaigns English-language, SEA geo.

Campaign 1 — Supplier-intent (30): [1688 supplier search], [find 1688 suppliers], [1688 english search], "1688 agent for shopee", "source from 1688", [chinese factory direct], "find suppliers in china", "china wholesale suppliers list", "1688 image search", "factory price sourcing china", "moq wholesale china", [sourcing platform for sellers], "how to find suppliers on 1688", "1688 without agent", "wholesale market china online", "cheap suppliers for dropshipping", "1688 to shopee", "1688 for lazada", "china sourcing tool", [supplier verification china], "real factory vs trading company", "minimum order quantity china suppliers", "product sourcing app", "find manufacturer by photo", "reverse image search supplier", "1688 price in usd", "bulk sourcing for ecommerce", "china wholesale site english", "best 1688 alternative", [source products for shopee].

Campaign 2 — 9.9 seasonal (20): "9.9 sale products to sell", "9.9 shopping day ideas", "what to sell 9.9 shopee", "9.9 voucher deals sourcing", "shopee 9.9 mega sale", "9.9 sale best sellers", "prepare for 9.9 sale", "9.9 product selection", "9 9 sale supplier", "flash sale stock china", "11.11 products to sell", "year end sale sourcing", [9.9 selling guide], "shopee mega day suppliers", "tiktok shop 9.9", "best products 9.9 2026", "stock up for 9.9", "wholesale 9.9 deals", "sourcing for mega sale", [9.9 sourcing free].

Campaign 3 — Competitor/alternative (15, generic-safe): "1688 alternative for shopee sellers", "better than 1688 for sellers", "1688 english alternative", "sourcing sites like 1688", "aliexpress vs 1688 for sellers", "taobao alternative english", "china sourcing tool for shopee", "1688 agent alternative", "find factories without 1688", "tool like 1688 image search", "1688 for beginners", "safest 1688 sourcing tool", "ai sourcing for ecommerce sellers", "1688 supplier finder app", [1688 alternative].

Negative keyword master list (~40), grouped: - Jobs/hiring: jobs, careers, salary, hiring, resume, internship, work from home, vacancy, recruitment, employee - B2C shopping: coupon, discount code, buy one, cheap phone case for me, amazon prime, free shipping code, retail, single item - Platform login/account: login, sign in, download app apk, install, password, account recovery, my orders - Courses/training: course, training, tutorial, learn, certification, class, coaching, how to start business - Freebies/scrapers: free coins, hack, mod apk, crack, template free excel, bot - Wrong region/intent: alibaba india, made-in-china india, nigeria, kenya, pakistan, craigslist, amazon fba usa

3 RSA assets (headlines ≤30 chars, descriptions ≤90 chars):

RSA-1 (Supplier-intent): H: "Find Real 1688 Suppliers" · "AI Sourcing Search" · "Source Direct From Factories" · "Price + MOQ In Seconds" · "Upload Photo, Get Suppliers" · "No Agent Needed" · "Verified Factories Only" · "Sourcing For SEA Sellers" · "Built For Shopee Sellers" · "See Factory Prices Now" · "Free Sourcing Search" · "Search In English" · "Compare Suppliers Fast" · "Stop Chasing Sellers" · "Try Free Until 9.9" | D: "Type a product or drop a photo — get source suppliers, price, MOQ and demand signals." · "Skip the language wall. Search source factories in English, in ~20 seconds." · "Every supplier scored on trading history and verification so you don't get burned." · "Free during the 9.9 sale period. No credit card. Find what to stock now."

RSA-2 (9.9 seasonal): H: "What To Stock For 9.9" · "9.9 Product Selection" · "Win The 9.9 Mega Sale" · "9.9 Supplier Search" · "Pick 9.9 Winners Fast" · "Free Until Sept 9" · "See 9.9 Hot Sellers" · "Stock Smart For 9.9" · "9.9 Sourcing Tool" · "Real Prices For 9.9" · "MOQ For 9.9 Orders" · "TikTok Shop 9.9 Picks" · "Shopee 9.9 Suppliers" · "Source Before 9.9" · "Verified 9.9 Suppliers" | D: "7 days to 9.9 — find the products and suppliers before your competitors do." · "Free AI sourcing search through the 9.9 sale. Photo to supplier list in seconds." · "Know price, MOQ and demand before you spend your 9.9 stock budget." · "Don't miss 9.9. See what's selling and who supplies it, right now."

RSA-3 (Competitor/alternative): H: "A Faster 1688 Alternative" · "Search Factories In English" · "Sourcing Without The Wall" · "Like 1688, But Easier" · "1688 For SEA Sellers" · "AI Supplier Finder" · "No Translator Needed" · "Find Suppliers Safely" · "Compare 1688 Suppliers" · "Skip The Agent Markup" · "Photo Search For Goods" · "Built For Shopee & Lazada" · "Verified Supplier List" · "Real Factory Prices" · "Try It Free Today" | D: "Still scrolling 1688 manually? One search returns ranked factories, prices and MOQ." · "A sourcing layer built for Southeast Asian sellers — English-first, image-ready." · "Cut agent markups and find source factories directly, with verification signals." · "Free until 9.9. Upload a photo, get real suppliers. No credit card required."

Ad schedule & bidding (first 7 days): Bid = Max Conversions with NO target CPA for days 1–3 (needs data), switch to tCPA set 20% above your observed CPA on day 4. Ad schedule: run 100% Mon–Sun but bid-adjust +30% on the 11:00–14:00 and 19:00–23:00 local seller-active windows (GMT+7/+8), −30% from 01:00–06:00. Device bid adjustment: +15% mobile, −20% desktop (this audience is mobile).

Location / language targeting rows:

TikTok Ads + Spark Ads

What to boost: Spark Ads on the 3 best organic videos from W2 — the raw screen-recording of "photo → 12 suppliers with prices", the "how I picked my 9.9 products" talking-head, and the "spot the fake factory" hook. Pull the authorization code from the video's TikTok app (Share → Ad → Get code, 60-day window) and attach to the ad so it keeps its organic comments, views, and creator handle — comments carry this product's social proof.

4 hook-first creative briefs:

1. Photo-to-factory. 0–3s hook (shot line): close-up thumb tapping a shirt photo into the search bar, on-screen text "one photo…". 3–15s body: results cascade — supplier, factory price, MOQ, demand bar climbing. On-screen text: "price · MOQ · demand in 20s". CTA end card: "Free until 9.9 — link in bio."
2. 9.9 panic. 0–3s: creator to camera, "9.9 is in a week and you still don't know what to stock?" 3–15s: shows typing "phone grip" → ranked hot suppliers. Text overlay: "stop guessing your 9.9 stock". CTA: "Find your 9.9 winner free."
3. Fake factory busted. 0–3s: "This 'factory' on a sourcing site is a reseller. Here's how to tell." 3–15s: highlights verification score, years trading vs a blank profile. Text: "verified or don't bother". CTA: "Source safe for free."
4. Time saved. 0–3s: split-screen, person drowning in 1688 tabs vs person who typed one query. 3–15s: "whole afternoon vs 20 seconds". Text: "I got my week back". CTA: "Try AI sourcing search."

Audience note: target interest in Online shopping, E-commerce, Dropshipping, Small business, plus creator-engagement audiences on Shopee/TikTok Shop selling content; age 20–40. Broad targeting works on TikTok because the algorithm reads the hook — keep interests as a hint only, lean on Spark social proof.

What NOT to do: do not run polished corporate ads (they're skipped in 0.5s); no logo sting at the open, no voiceover that sounds like a TVC, no link-in-bio-less traffic ads, no more than one CTA, don't repurpose static Meta feeds into feed slots, and never buy Spark authorization from an account that doesn't match the handle you want viewers to trust.

Retargeting & Closed Loop

Audiences to build (all with the paid→owned handoff in mind):

Frequency caps: total paid exposure per user across Meta+TikTok ≤ 5/7d; retargeting sets mutually exclude each other and exclude all prospecting converters (registered users excluded from every retarget set). Signup-but-no-search is the priority loop — it's the cheapest revenue and the most likely to churn silent.

Paid vs owned handoff rule (no double-hit): On CompleteRegistration (or Instant Form submit), a person is immediately added to the "registered" exclusion on ALL Meta/TikTok/Google audiences — paid stops within 1 hour. Ownership moves to email + WhatsApp: welcome + "start your first search" (WhatsApp for VN/MY/ID, email for TH/PH), then a day-2 "your 9.9 category report" nudge. Paid only re-engages a registered-but-dormant user through the signup-without-first-search loop after 3 days of no search. That loop is the ONLY place paid touches a registered user. This guarantees no one gets a cold ad and an owned message for the same action.

Localization: Ad Copy in ID / TH / VI

Top-3 priority angles: (a) 9.9 selection anxiety, (b) price/MOQ transparency, (f) free-during-9.9. Rewritten in marketing register, not literal.

Back-translations (sanity check): ID-a "9.9 is just a few days away. Still confused about what to stock? Just type or send a photo — straight to supplier price, MOQ, and what's selling hot." TH-b "Two numbers that decide profit: price and MOQ. See both immediately from the factory, no long haggling, no language worry." VI-f "Free until 9.9. Upload a product photo — find the real original factory, see price, MOQ and how hot it sells. No credit card needed."

Which markets need which localization: ID → Bahasa Indonesia mandatory (English-only underperforms badly). TH → Thai mandatory, script and tone both matter. VN → Vietnamese mandatory, but pair with the Zalo/WhatsApp landing flow. MY → English + Bahasa both convert (large English-speaking seller base). PH → English is fine, no translation needed.

Native-reviewer QA step: before any localized creative spends >$20, have one native seller-side reviewer (from the W2 Facebook group in that country) confirm the copy reads like a local seller talking, not like a translation, and flag any awkward transliteration of "supplier/MOQ/1688" and any date/number formatting (9.9 vs 9/9). Log the reviewer's name and green-light date in the campaign notes.

Launch-Week Ops: Daily Paid Routine

1. Mon Sep 8 (go-live, 09:00 local). Confirm Pixel+CAPI firing, Event Match Quality ≥6, Aggregated Event order correct, exclusions live. Push $60/day. Check only that spend started and no ad was rejected. Notify: paid operator + growth lead ("LIVE, all green/red").
2. Tue Sep 9 (9.9 DAY — free promo cutoff). Morning: verify all "free until 9.9" creatives are still truthful — flip to "last day of free" messaging or prepare the post-9.9 angle by afternoon. This is a high-intent day: +20% budget on Sales-optimized Meta and 9.9 Google keywords for the day. Notify: whole team, log the 9.9 spike in the daily sheet.
3. Wed Sep 10. First real read: per-ad-set CPA vs target, thumb-stop on TikTok, LP→signup CVR. Kill any creative thumb-stop <25% (TikTok) or CTR <0.8% (Meta). No budget moves — learning phase.
4. Thu Sep 11. Exited-learning ad sets: turn on tCPA/tROAS where ≥15 conversions. Check FirstSupplierSearch completion rate (signup→first-search) — if <40%, the problem is onboarding, not the ad; flag to product. Notify: product lead if activation is low.
5. Fri Sep 12. Retargeting loop should now have an audience pool. Confirm handoff rule: pull registered users' paid impressions to ~0. Review negative-keyword search terms in Google, add new negatives.
6. Sat Sep 13. Mid-week scaling call: shift test-reserve budget to the single best angle. If blended CPA ≤$3.00 for 3 days, pre-authorize the Wk2 $100/day ramp. Notify: growth lead with the go/no-go.
7. Sun Sep 14. Week retro numbers into the dashboard: spend, signups, CPA by channel, first-search activation, retarget ROI. Draft Monday's Wk4/11.11 pre-heat keyword adds. Freeze changes before Monday.

Daily morning check (all days): yesterday spend, blended CPA, any rejected/disapproved ads, budget-capped campaigns, Event Match Quality, and whether the 9.9-messaging deadline logic is still correct.

Measurement Guardrails

The 6 paid metrics in the daily review (target value + signal):

The 3 traps to avoid: 1. Judging on CTR/CPM instead of signup CPA and activation. A cheap-CPM ad set that produces signups who never run a first search is a loss; only FirstSupplierSearch proves the channel brings real sellers. Read CPA and activation together. 2. Mixing/layering edits that reset the learning phase. Editing budget, audience, or creative on an ad set still in learning (or <50 conversions) resets it and re-spends; batch changes, scale in ≤25% steps, and never judge a set before it exits learning or hits day 5. 3. Ignoring the funnel gap between signup and first search. A low CPA with a weak signup→search rate means you're buying bad traffic or have broken onboarding — optimizing the ad won't fix a product-drop. Watch CVR-to-first-search as a first-class guardrail, and when it's low, alert product rather than blaming the ads.