# Your first ads on Meta and Apple Search (/docs/playbooks/first-ads-meta-apple)
## Who this is for [#who-this-is-for]
You have a product that works, some organic traction, and you're ready to try paid ads. You've never run paid ads before, or you've tried and got mediocre results and want a cleaner restart. You have $500-$3000 of discretionary budget for the experiment.
This playbook assumes a consumer-facing mobile app. For web-first or B2B, most of the tactics still apply but the measurement layer changes — you won't need the SDK, and you'll measure on-site conversions instead.
## When to use this [#when-to-use-this]
* Retention is working: users who sign up are still around in week 2 at a reasonable rate for your vertical
* You have at least 3-5 known-good organic creatives (posts that outperformed your median by 2-3x)
* You have budget for at least 2-3 weeks of learning before you call results
* You understand you might spend the whole budget and learn nothing, and you're OK with that
If retention isn't working, don't run paid yet. You'll just burn money acquiring users who won't stick. Read [retention-driven growth](/docs/playbooks/retention-driven-growth) first.
## The honest setup [#the-honest-setup]
Two things people don't tell you about first-time paid ads:
**1. The first $500-$1500 is tuition, not acquisition.** You're paying the platforms to teach their algorithm who your customer is. The users you get in this phase might not be the right users. Treat this as a learning investment.
**2. Your best organic creatives are your starting ad creatives.** Almost nobody writes a net-new ad that beats a proven organic post. The ads that work are usually reruns of content that worked organically, sometimes with a tighter call-to-action.
## Why Meta + Apple Search (and not Google) [#why-meta--apple-search-and-not-google]
For mobile consumer apps in 2026:
* **Meta (Facebook + Instagram)** is the workhorse for broad acquisition. Lower CPA than most competitors, works for almost any vertical.
* **Apple Search Ads (ASA)** is the most efficient channel for iOS-specific apps with keyword intent. Someone searching the App Store for "meditation app" is intent-heavy.
* **Google Ads (UAC)** is powerful but more complex, and the attribution is noisier for mobile. Worth it after you've got Meta + ASA working.
* **TikTok Ads** is worth adding at scale (past $100/day) but is harder as your first channel because creative requirements are steeper.
For a first ads run, Meta and ASA are the combination. Meta teaches you about creative; ASA teaches you about intent.
## Set up measurement before you set up ads [#set-up-measurement-before-you-set-up-ads]
This is the order people get wrong. They launch ads, realize they can't measure conversions, and end up scaling based on click numbers. Don't do that.
**Install the [Layers SDK](/docs/sdk)** in your app. Quickstart takes about 10 minutes per platform. You need at minimum:
* `install` event (first open)
* One activation event (signup, tutorial complete, first key action — pick the one that signals real usage)
* `purchase` event if you have paid conversion, with `revenue` and `currency`
**Set up CAPI for Meta.** Layers SDK + CAPI relay handles this server-side. If you're running Layers in BYO ad account mode, follow [the Meta BYO setup](/docs/paid-media/meta-setup#byo-mode). The one-time setup pays off: CAPI recovers 15-30% of conversions that iOS 14.5+ privacy changes made invisible to Pixel alone.
**Verify test events arrive.** In Meta Events Manager, check that your test events are showing up with the test event code. In the [SDK event inspector](/docs/sdk/event-inspector), confirm events fire from the client.
If this step takes you two days, great. The worst outcome is launching ads while measurement is broken and making decisions on garbage data for a month.
### What failure looks like [#what-failure-looks-like]
You launch ads, realize your conversion event isn't firing, fix it two weeks later. Now your two weeks of data is useless because Meta's algorithm optimized for clicks instead of installs. Start over with budget burned.
## Pick your first campaigns [#pick-your-first-campaigns]
Start narrow. You want two, maybe three campaigns max.
### Campaign 1: Meta — App Install (broad) [#campaign-1-meta--app-install-broad]
* **Objective:** App Install or Conversions (if you're measuring a deeper event)
* **Budget:** $30-$50 per day. Do not start lower — Meta's algorithm under-performs on sub-$20/day adsets in 2026.
* **Audience:** Broad. Don't over-target. One audience, maybe a country and an age range, nothing fancy.
* **Creative:** 3-5 of your best organic videos, each as its own ad
* **Placements:** Automatic. Resist the urge to hand-pick.
### Campaign 2: Apple Search Ads — Search Results [#campaign-2-apple-search-ads--search-results]
* **Campaign type:** Search Results (not Search Tab, not Today Tab — start with the highest-intent placement)
* **Budget:** $20-$30 per day
* **Keywords:** Start with 10-20 high-intent keywords relevant to your app. Include your brand name, your category ("meditation app"), and a handful of competitor names.
* **Bid strategy:** Start with Apple's suggested bids, adjust after 7 days
[Apple Search Ads setup in Layers](/docs/paid-media/apple-setup) walks through the full connection flow. ASA is agency-only in Layers (Apple doesn't support self-serve delegation for this kind of tool), so you'll be configuring through Layers' agency mode.
### Campaign 3 (optional): Meta — Retargeting [#campaign-3-optional-meta--retargeting]
Skip this for the first two weeks. You need install data first. Once you have 1,000+ installs, you can run a second Meta campaign retargeting people who installed but didn't complete activation. That campaign usually outperforms the acquisition campaign on CPA by 2-3x.
## Budget math — what to expect [#budget-math--what-to-expect]
Brutally honest numbers for a first run:
| Budget | What you get | What you learn |
| ------ | -------------------------------------- | ------------------------------------------------------------- |
| $500 | 50-300 installs, depending on vertical | Whether your creative is fundamentally weak or strong |
| $1,500 | 200-1,000 installs, a few conversions | Which creative angle works; whether your onboarding is a leak |
| $3,000 | Statistical confidence on CPA | Whether the unit economics could work at scale |
If your product has a clear revenue event and retention, you can calculate target CPA as:
```
target_CPA = payback_LTV × target_margin
```
Where `payback_LTV` is how much a user pays you over the payback window you can tolerate (typically 3-6 months for consumer apps). If your 6-month LTV is $20 and you want 50% margin, your target CPA is $10.
If you don't know your LTV yet, start by learning your CPI (cost per install) and your install-to-activation rate. You need both before scaling.
## The first two weeks: what to watch, what to ignore [#the-first-two-weeks-what-to-watch-what-to-ignore]
The first 7-14 days are Meta's learning phase. During this window:
**Ignore:**
* Daily CPA fluctuations (noisy at low volume)
* Individual ad performance rankings (not enough data)
* The urge to turn things off
**Watch:**
* Whether Meta is spending your budget (if daily spend is way below target, your audience is too narrow or your bids are too low)
* Whether ASA is exiting learning on keywords (usually 3-5 days)
* Directional CPA — is it in the same order of magnitude as your target?
After day 14:
* Kill ads that spent $40-$50 with no conversions (statistical confidence of badness)
* Raise budget 20-25% on ad sets with CPA below target
* Do NOT raise budget by more than 25% in one move. Meta's learning phase resets on big budget changes.
This is roughly what the Layers [paid-media optimizer](/docs/paid-media/optimizer-decisions) does automatically. If you're using Layers in agency mode, you can watch it happen in the optimizer activity log. If you're running BYO, you'll do it manually.
### What failure looks like [#what-failure-looks-like-1]
Day 4: "This isn't working, I'm turning it off." You killed the campaigns before Meta's learning phase completed. Nothing learned, budget half-spent, no insight. Fix: commit to a 14-day runway when you start, and only kill individual ads (not the whole campaign) during the learning window.
## The first conversion — what it tells you [#the-first-conversion--what-it-tells-you]
When your first paid conversion comes in, you learn three things:
1. **The funnel works end-to-end.** Install → activation → conversion all fire. Even one is enough to prove this.
2. **The cost is measurable.** Even if it's $80 for a $10 product, you now have a number to work on.
3. **The creative has something.** Of your 3-5 ads, one probably drove that conversion. That ad is your first signal.
What the first conversion does NOT tell you:
* That your strategy works (one data point)
* What your actual CPA will be (too noisy)
* Which of your creatives is the "winner" (need 10+ conversions per creative before ranking)
Keep going. Don't extrapolate from a single conversion.
## Week 3-4: when to scale, when to hold [#week-3-4-when-to-scale-when-to-hold]
By weeks 3-4, you should have enough data to answer:
* Is my CPA in an acceptable range? (within 2-3x of target = promising; within 1.5x = scaling territory)
* Do I have at least one ad that's clearly outperforming?
* Is the install-to-activation rate holding up for paid users vs organic users? (If paid users activate at half the rate of organic, you have a creative-audience fit problem)
### Scale signal [#scale-signal]
All three below are true:
* CPA within target range on at least one ad set for 7+ days
* At least 30-50 conversions captured across the account
* Activation rate for paid users within 20% of organic users
If all three: raise budget 25% on the winning ad set. Wait 5-7 days. Raise another 25% if the CPA held. [The scaling playbook](/docs/playbooks/scaling-ad-budget) covers what comes next.
### Hold signal [#hold-signal]
* CPA is 2-3x target but still improving week over week
* One or two creatives showing promise but not enough conversions to confirm
Hold the current budget for another 1-2 weeks. Don't scale yet; don't kill. You're still in learning.
### Kill signal [#kill-signal]
* CPA 5-10x target with no improvement trend after 3 weeks
* Paid users activate at less than half the rate of organic users (you're buying the wrong users)
* You've spent $2000+ with no conversions at all
Kill the campaign. Diagnose before relaunching. Most "kill" outcomes are creative problems, not platform problems. Rotate in [fresh creative](/docs/playbooks/creative-refresh-cadence) and rerun.
## The creative rotation problem [#the-creative-rotation-problem]
Even if your first ads work, they'll fatigue. The same creative shown to the same audience stops performing after 7-14 days of real spend. This is the single most common reason first paid-ads runs die in month 2.
The fix is a creative refresh cadence, not better targeting. See [creative refresh cadence](/docs/playbooks/creative-refresh-cadence) for the mechanics. The short version: have 2-3 new creatives ready to swap in every two weeks, permanently.
This is where a tool like Layers earns its keep. Manually writing and producing fresh ad creatives every two weeks is a part-time job. [Layers' creative studio](/docs/paid-media/creative-studio) and [UGC pipeline](/docs/ugc) can feed fresh variants into the creative pool, and the [optimizer](/docs/paid-media/optimizer-decisions) handles rotation automatically. Without some version of this, you'll stall out.
## What "success" looks like for first paid ads [#what-success-looks-like-for-first-paid-ads]
A realistic success outcome for a first 30-day paid run:
* CPA within 2x of target by end of month
* 1-2 ads identified as consistent performers
* At least 200-500 attributed installs
* A clear sense of what your real CPA ceiling is
* A list of 3-5 things you'd do differently in month 2
What it does NOT look like:
* 5x ROAS in week 1
* A viral ad that made your day
* "Set and forget" performance
* Confidence that your strategy will scale to $10k/day
If your month 1 results are "measurable, directional, and modest," that's the win. Month 2 is where things usually click.
## What's next [#whats-next]