The Complete Guide to Meta Ads for GrubList

Use the Meta SDK directly — not an MMP (Mobile Measurement Partner) like Adjust or AppsFlyer. At GrubList's stage, the SDK is free, avoids third-party complexity, and gives you everything you need.

Events to Track for GrubList

Don't replicate your product analytics in Meta. Only track 3-5 events that matter for user acquisition:

Why $10K minimum? The Meta algorithm needs conversion volume to learn. If you spend $3K/month and your cost per trial is $20, you're getting ~150 trials total — barely enough signal for Meta to figure out who to target. You need the 50 events/ad set/week threshold.

Budget Math for GrubList

Creative Production Budget

Hire someone for your first ads. If creative quality is an unknown variable AND the channel is an unknown variable, you'll never know which one failed. Budget $5K for a freelancer or small agency to produce:

• 3-4 UGC-style videos (30-60 seconds)
• 3-4 static ads (for Facebook feed)
• Variations with different hooks

Initial Structure

Why NOT Advantage+?

Advantage+ (ASC+) is Meta's AI-driven campaign type and works well at scale. But it has a critical limitation: age exclusions require account-level settings, which is complicated. Manual campaigns let you exclude under-25 at the ad set level, which is essential for a subscription app like GrubList.

Test Advantage+ later when you have data and volume. They can buy completely different audiences and placements — worth testing, but not simultaneously at small budgets.

Your creative IS your targeting. The ad itself determines who sees it. For GrubList, your ads need to speak directly to the person who's exhausted by the "what's for dinner?" question.

UGC Video Concepts (for Reels)

Static Ad Concepts (for Facebook Feed)

Three Rules of App Creative Testing

Rule 1: Test on Android first. If GrubList is on both platforms, run creative tests on Android. It's cheaper, tracking is better (no SKAN headaches), and results port to iOS ~90% of the time.

Rule 2: Keep a separate testing campaign. Never test new creative inside your performing BAU (business-as-usual) campaign. New ads inject volatility. Run a dedicated testing campaign with its own budget.

Rule 3: Separate by creative type. In your testing campaign, create separate ad sets:

Testing Cadence

• Test 3-5 new concepts per week (not variations of the same concept)
• Give each creative at least 2x your target CPA in spend before killing it
• Winners (matching target CPA after one budget increase) graduate to BAU campaign
• Use AEM attribution for testing — results come in real-time instead of 2-3 day SKAN delay

Here's what happens without intervention:

1. Meta finds that 20-year-olds on Instagram Reels are cheap to acquire
2. Your cost per trial looks great — $12 instead of $20
3. You celebrate and increase budget
4. Two weeks later, you check RevenueCat: trial-to-paid conversion is 15% instead of 40%
5. Your actual CPA (cost per paying user) is $80, not $30

How to Fix This for GrubList

Step 1: Add an age question to your onboarding. "What's your age range?" — 18-24, 25-34, 35-44, 45+

Step 2: In your product analytics, segment trial-to-paid conversion by age group. You'll likely find that 35-44 year old parents convert at 2-3x the rate of 18-24 year olds.

Step 3: Use campaign structure to guide Meta toward valuable audiences:

• UGC video ad set will naturally skew young (Reels audience)
• Static ad set will naturally skew older (Facebook Feed audience)
• If statics have 50% higher cost per trial but your data shows 2x trial-to-paid conversion, statics are the better investment

Step 4: In Manual campaigns, you can exclude under-25 at the ad set level. Consider this once you have data confirming the age/conversion gap.

Recommendation for GrubList

Start with AEM — it's the default and gives you faster data, which is critical during your testing phase. You don't want to wait 3 days to learn if a creative is working.

Later, once you have a BAU campaign running, test switching it to SKAN and compare audience quality. Some advertisers find SKAN drives better audiences. You can run both simultaneously — they may buy completely different users.

When you turn on Meta ads for GrubList, watch these numbers in parallel:

What to Track

• RevenueCat / product analytics: Total trials, trial-to-paid conversion, revenue — broken down by onboarding survey age group
• App Store Connect: Watch "Search" traffic — this includes people who saw your Meta ad and searched for "GrubList" in the App Store
• Meta Ads Manager: Reported trials (will under-report — that's expected)
• Baseline comparison: What were your organic metrics before Meta? The difference = Meta's actual impact

The Blended Equation

If you're just starting, skip lookalikes. Go broad and let the algorithm learn from your creatives. Lookalikes require significant existing data to be effective.

When Lookalikes Make Sense

• You have 10,000+ user entries to upload (device IDs, IDFAs, emails)
• Meta can typically match ~20% of your list, so you need 10K to get a ~2K matched audience
• Email match rates are terrible (people use different emails for Facebook vs your app)

GrubList Lookalike Strategy

Once you have 10K+ paying subscribers in your database:

1. Export subscriber list from RevenueCat (IDFAs where available + emails)
2. Upload as Custom Audience to Meta
3. Create 1% lookalike (narrowest, highest quality)
4. Test alongside your broad targeting — run both and compare

When people see your Meta ad, many will search the App Store instead of clicking. That means your App Store listing IS a landing page for your paid ads.

Optimization for Meta Traffic

• Match your winning ad creative to your App Store screenshots. If your best ad shows a "receipt comparison" format, make that your first screenshot. People need to feel they landed in the right place.
• Make your first screenshot communicate GrubList's core benefit in 2 seconds. Marcus Burke noted that apps where you can't instantly tell what they do lose conversions.
• Add contrast between screenshots. Don't make every screenshot look the same — varied colors and layouts increase engagement.
• A/B test your screenshots — making your main keyword bold and prominent increased conversions by 30% for one of Marcus's clients.

Pre-Launch (Before Ads)

• Meta SDK installed and tested
• Trial Start event firing correctly
• Onboarding Complete event set up
• First List Created (qualification event) tracked
• Age question added to onboarding survey
• RevenueCat/product analytics baseline established (2+ weeks)
• App Store screenshots optimized (match ad creative)
• $10K media budget allocated for month 1
• 3-5 ad creatives produced (mix of UGC video + static)

Campaign Launch

• 1 Manual campaign created with App Install objective
• 1 ad set, broad targeting, Trial Start optimization
• AEM attribution selected (default)
• 3-5 creatives loaded into ad set
• Daily budget set ($300-350/day)
• Android creative test campaign set up (if dual platform)

Week 1-2 Monitoring

• Check cost per trial daily — is it $15-25 range?
• Check placement breakdown — where are ads delivering?
• Check age breakdown in onboarding data — who's signing up?
• Compare organic baseline to current — is there search uplift?
• Kill any creative spending 2x CPA with zero trials

Month 1 Review

• Calculate blended CPA (Meta reported + organic uplift)
• Check trial-to-paid conversion by age group
• Decide: separate static vs video into 2 ad sets?
• Set up dedicated creative testing campaign
• Decision: continue, scale, or pause Meta based on blended ROAS