How to Launch Your First Meta Ads Campaign From Scratch

Most first Meta campaigns fail not because of the algorithm, but because of setup decisions made in the first ten minutes. Wrong objective. Mismatched audience. Creative that qualifies nobody. Budget so low the algorithm never escapes the learning phase. The result is a campaign that spends money, produces noise, and tells you nothing useful.

TL;DR: To launch meta ads from scratch, you need five things in place before you hit publish: a verified Business Suite account, a firing Pixel (plus CAPI), a clear campaign objective matched to your funnel stage, creative with a defined cold-audience hook, and a daily budget high enough for the algorithm to learn. Get those five right and the campaign will give you real signal within 7–10 days.

This guide walks through every decision in sequence — not as a UI walkthrough, but as a reasoning guide. Each section explains what you are deciding and why it matters when you launch meta ads from scratch, so you can adapt when Meta changes the interface next quarter.

Why Most First Campaigns Fail Before They Run

The failure mode is almost always structural, not creative. Advertisers set a campaign objective that doesn't match their funnel stage (Traffic for a store that needs purchases) then wonder why CPAs are astronomical. They build three interest-stack audiences that overlap, each underfunded, none of them exiting the learning phase. They spend €15/day, see a CPA of €200 on day two, and turn the campaign off before any real data exists.

The Meta algorithm needs events to learn from. Specifically, it needs roughly 50 optimisation events per ad set per week to exit the learning phase and shift from exploration to exploitation. Below that threshold, it is guessing. Every structural choice you make before launch either helps the algorithm get those 50 events faster, or makes it harder.

Start with that frame: your job at setup is to create the conditions for the algorithm to learn.

What You Need Before You Open Ads Manager

Four things must be in place before you create a single campaign.

1. Meta Business Suite and a verified ad account. Go to business.facebook.com, create a Business Suite account, add your Facebook Page and Instagram profile, and create an ad account. Verify your business domain under Business Settings > Brand Safety > Domains. Without domain verification, some conversion tracking features are restricted.

2. The Meta Pixel and Conversions API installed. The Meta Pixel is a JavaScript snippet that fires browser-side events. Install the base code site-wide, then configure standard events: ViewContent on product pages, AddToCart, InitiateCheckout, Purchase. The Conversions API (CAPI) fires the same events server-side, giving the algorithm deduplication signal it needs after iOS 14 gutted browser-side data. If you skip CAPI, expect 20–40% signal loss on iOS devices. The Meta Pixel guide covers the full Pixel + CAPI + AEM stack if you need implementation detail.

3. A defined offer. Know exactly what you are selling and to whom. A vague product with no clear hook produces vague ads. Before writing a single line of copy, write one sentence: "This ad is for [specific person] who has [specific problem] and will get [specific outcome]." The offer clarity problem is upstream of every creative problem.

4. Three to five creative assets ready. You will need them at launch. See the creative section below for what to build.

Campaign Objective — The Decision That Shapes Everything Else

The campaign objective tells Meta's algorithm what signal to optimise toward. Pick the wrong objective and the algorithm will get very good at something you don't want.

Sales (Conversions): Use this for ecommerce and direct response. Set the optimisation event to Purchase if your Pixel has 50+ purchase events in the last 30 days. If not, use AddToCart or InitiateCheckout as a proxy and move to Purchase optimisation once you have enough data. The CBO vs ABO guide explains how budget allocation interacts with objective choice.

Leads: Use for B2B, service businesses, or any campaign where the conversion is a form fill. Optimise toward a qualified lead event, not a raw form view.

Traffic: Use only if you have zero Pixel data and need initial event accumulation. Stop using it once you have 50+ conversion events — at that point, Traffic optimisation has nothing to offer.

Awareness / Reach: Not for first campaigns. These are brand objectives with no direct conversion optimisation. Useful once you have a working conversion campaign to complement.

The rule: always optimise toward the event that matches your actual business goal, unless the Pixel doesn't have enough of that event to learn from — in which case, pick the event one step up the funnel and migrate down when the data is there.

Audience Setup: Cold, Warm, and Broad

For a first campaign, you will likely be running to a cold audience — people who have never interacted with your brand. The temptation is to layer in interest targeting to pre-qualify. Resist it.

Interest stacks narrow the audience pool before the algorithm has enough data to find signal inside it. They raise CPM (fewer people competing for fewer impressions), slow the learning phase, and make it harder to diagnose results — if CPAs are high, you can't tell whether the creative is wrong or whether the interests just excluded the right people.

Broad targeting: Set location, a wide age range (18–55 unless you have strong evidence to narrow), and leave interests blank. Let Advantage+ audience or the algorithm find the signal in your creative and Pixel data. The broad targeting guide explains why this works better than it should, especially post-Andromeda.

Advantage+ audience: In campaign setups that offer it, Advantage+ audience replaces manual detailed targeting with algorithmic audience expansion. For a first campaign with limited Pixel history, this is generally the better starting point.

Retargeting layer: Separately, build a custom audience of your site visitors (all visitors, last 30 days), add-to-carts (last 14 days), and video viewers (if you have video content). Keep retargeting in a separate campaign with a smaller budget — don't mix cold and warm in the same ad set. The retargeting guide covers segmentation logic in full.

Lookalike audience: Once you have 100+ purchase events in your Pixel, a 1–3% lookalike audience of purchasers is worth testing. Not for launch — you need the seed data first.

Audience segmentation strategy determines your CPM and how quickly the algorithm finds signal. Keep it simple at launch.

Budget: Start Low and Learn Fast

Budget sizing is a learning problem, not a spending problem. The question is: what daily budget gets you to 50 optimisation events per week without wasting money on a campaign that hasn't converged?

For most ecommerce campaigns with a purchase CPA in the €30–€80 range, €40–€60/day per ad set gives you roughly 3–5 purchases per week at the high end of expected CPA. That's enough to clear the learning threshold in 10–14 days.

If your product has a €20 CPA target and you set a €20/day budget, you'll need exactly one purchase per day to be profitable — and the algorithm will never exit learning because a single event per day is not enough signal. Start at 2–3x your target CPA as a daily budget floor, then scale once the campaign has exited learning.

Use CBO (Campaign Budget Optimisation) for first campaigns. Set the budget at the campaign level and let Meta allocate across ad sets. ABO gives you more control but requires more data to optimise manually.

Use the ad budget planner to estimate how long your budget will take to clear the learning phase based on your expected CPA and daily spend.

Creative: What to Build for a First Test

Creative is where most first campaigns actually fail. Not targeting, not budget — creative. The algorithm can find the right people if the creative signals who it's for. A generic ad signals nothing.

For a first cold-traffic test, build three to five creative variants across two formats:

Direct static (1–2 variants): Product image or clean lifestyle shot, price or offer visible, clear CTA button or text. This is a Direct ad — it explicitly sells. Works for warm audiences and bottom-funnel cold audiences. Use 1:1 or 4:5 ratio for feed placements.

Native / UGC-style (1–2 variants): Looks like organic content. A person talking to camera, a before/after, a testimonial screenshot. No visible brand logos in the first three seconds. This works better for cold audiences who don't know you yet. See the UGC ads guide for what makes the format work.

Short video (optional, 1 variant): Under 15 seconds, hook in the first 3 seconds, mobile-first 9:16. If you have the asset, test it. Video tends to win on cold audiences in competitive categories. See the video ads guide.

All creative must answer one question in the first three seconds: what is this and why should I care? The hook rate measures how many people watched past the 3-second mark. Aim for 25%+ on your best performer.

Before building creative, spend 20 minutes in the AdLibrary ad detail view searching for competitors in your category. Look at what formats they are running, how long those ads have been active (longer = working), and what hooks they are using. Creative research before you build is not optional — it takes 20 minutes and saves you a week of bad tests.

If you need a structured approach to developing your creative brief before production, the creative brief template gives you the research-first method.

Ad Set and Ad Configuration

With objective, audience, and budget set at the campaign level, the ad set is where you configure placements and scheduling.

Placements: Use Advantage+ placements (Meta's automatic placement) for a first campaign. It distributes across Facebook Feed, Instagram Feed, Reels, Stories, Audience Network, and Messenger automatically. If you want to isolate Reels or Stories performance, create a separate ad set with manual placement. Don't restrict to a single placement at launch — you need the distribution breadth for the algorithm to find signal.

Ad set budget (if using ABO): €30–€50 per ad set per day, as above. With three ad sets, that's €90–€150/day total — appropriate for a mid-size first campaign.

Ad level: Create 3–5 ads per ad set using your creative variants. Each ad is a distinct creative: different image or video, different primary text, different headline. You are not running multivariate testing — you are giving the algorithm real creative variance to find the winner among. The dynamic creative feature can do some of this automatically if you prefer.

Ad copy: Primary text should open with the hook, not the brand name. Facebook primary text is 125 characters before "See more" truncation. Get the value proposition in those first 125 characters. Headline (the line below the image) should be direct: product name + primary benefit, or a specific offer. See the ad copy guide for cold-traffic copy principles and the ad headline guide for headline frameworks.

UTM parameters: Add UTM tags to every ad URL. utm_source=facebook&utm_medium=paid&utm_campaign=launch-v1&utm_content=creative-name. This lets you cross-reference Meta's reporting with your own analytics. Meta's attribution and your analytics will disagree — that's normal. Both views matter.

Publishing and the Learning Phase

Before you hit publish, run through this checklist:

• Pixel fires on checkout confirmation page (verify in Meta Events Manager)
• Campaign objective matches your conversion event
• Budget is set at a level that will generate 50+ events per week
• Creative is approved and has correct aspect ratios for selected placements
• UTM parameters are on all destination URLs
• Payment method is verified on the ad account

Once you publish, the campaign enters the learning phase. This is the window where Meta's algorithm runs exploration — it tries different people, times, and placements to find where your conversion event fires most reliably. It needs roughly 50 optimisation events to exit this phase. During the learning phase:

• Do NOT edit the budget by more than 20%
• Do NOT change the targeting
• Do NOT pause and restart the campaign
• Do NOT add or remove creatives from an ad set

Any of these actions resets the learning phase. If your campaign is delivering and spending, leave it alone for 7 days. Check delivery daily. If there is a zero-delivery issue (spend not starting), see the Facebook ads not delivering diagnostic.

The learning phase calculator estimates how many days your specific budget and CPA will take to exit learning.

Reading Your First Results

After the learning phase exits, you have a real baseline. Here is how to read it.

CPA vs target: Compare your achieved CPA against your break-even threshold. Use the break-even ROAS calculator to find your floor. If CPA is within 1.5x of target at learning-phase exit, you have signal worth scaling. Between 1.5x and 2.5x, the creative needs work. Above 3x, something structural is wrong — objective, audience, or offer.

ROAS: Compare against your minimum viable ROAS. Use the ROAS calculator to establish thresholds before you start, not after. Post-launch, ROAS below 1.0 means you are losing money on every euro of spend.

Hook rate: How many people watched 3 seconds of your video, or paused on your static? Hook rate below 15% means your creative doesn't stop the scroll — it's a creative problem. Above 25% means you have attention but may have a conversion problem if CPA is still high.

Thumb-stop ratio: Similar to hook rate, tracks 2-second video views. Target 20%+. See the thumb-stop ratio guide for benchmarks.

Creative-level breakdown: Sort ads by CPA ascending. The lowest-CPA creative is your winner. Kill the highest-CPA creative if it's spent 3x target CPA with no conversion. Leave the middle performers running until they have enough spend data to evaluate (minimum 3x target CPA in spend).

Scale by increasing the budget on the winning ad set by 15–20% every 3–4 days. More aggressive increases reset learning. The spend-scaling roadmap walks through the full phase-gate method from launch to €500k/month.

Frequently Asked Questions

How much budget do I need to launch Meta ads from scratch?

A minimum of €30–€50 per day per ad set gives the algorithm enough events to exit the learning phase within 7–10 days. Below that, the campaign can stall in learning indefinitely. Start with one ad set, one objective, and 3–5 creative variants.

Which campaign objective should I pick for my first Meta ads campaign?

For ecommerce, choose Sales (Conversions) and optimise for Purchase. If your Pixel has fewer than 50 purchase events, optimise for AddToCart or InitiateCheckout first to accumulate signal. Use Leads for B2B or service businesses. Avoid Traffic unless you have no Pixel and just want data.

Do I need to install the Meta Pixel before running my first campaign?

Yes. Without the Pixel (or Conversions API), Meta cannot optimise toward your actual business goal — it will only optimise for clicks. Install the Pixel base code, configure at minimum the Purchase event, and add CAPI for server-side data to compensate for browser-side signal loss from ad blockers and iOS. The full implementation guide is in the Meta Pixel post and Conversions API post.

What targeting should I use for a first Meta ads campaign?

Start broad. Set location and a wide age range. Leave interest and behaviour targeting empty or use Advantage+ audience. Interest stacks narrow the pool before the algorithm has signal, which extends the learning phase and raises CPM. Broad targeting works best when the creative does the qualifying.

How do I know if my first Meta ads campaign is working?

Wait for the learning phase to exit — typically 50 optimisation events, roughly 7–10 days. Then compare CPA against your break-even threshold, ROAS against your minimum target, hook rate (aim >25%), and thumb-stop ratio (aim >20%). If CPA is within 2x of target at exit, the campaign has signal worth scaling. If it is 3x or more, the issue is almost always creative.

Launch the campaign. Read what it tells you. Iterate. That's the whole system for how to launch meta ads from scratch.

Three structural mistakes appear in nearly every failing first meta ads campaign. According to Meta's own Business Help Center, the learning phase requires roughly 50 optimisation events per week — and these mistakes reliably prevent that threshold.

1. Optimising for the wrong event. If your Pixel has 10 purchases in the last 30 days and you optimise for Purchase, your meta ads campaign has almost no signal to work with. Fix: optimise for AddToCart or InitiateCheckout first, then migrate to Purchase once data accumulates. The Facebook ad campaign structure guide explains when and how to migrate.

2. Editing during the learning phase. The most common self-inflicted wound. Someone checks meta ads results at day 3, panics at a €90 CPA, adjusts the audience, and resets learning. The CPA at day 3 is meaningless — exploration mode, not exploitation. Read the Meta campaign planning guide for a full breakdown of what each meta ads phase actually means.

3. Under-investing in creative. Spending €50/day on one creative variant is running an A/B test with one variant. Three to five variants at launch is the minimum for real meta ads data. The campaign structure guide explains how to organise them without polluting ad set data.

For a deeper read on what kills Meta campaign performance, the diagnostic post covers 22 root causes with a decision table. For specifics on ad set optimisation mechanics, the bid cap post explains when manual bidding helps meta ads learning and when it stalls it.

Research Before You Launch Meta Ads: Using AdLibrary

Most first-time advertisers build meta ads creative in a vacuum — writing copy based on what sounds good, picking images based on what looks professional, guessing at hooks without knowing what the category is already saturated with. Thirty minutes of competitor research before launch changes both the brief and the results.

Open AdLibrary's platform filters and search for 3–5 competitors. Filter by Meta. Sort by longest-running ads — unprofitable meta ads get turned off, so long-running means profitable. Note format distribution (static vs video vs carousel), primary text patterns, and what offer structure is leading. This is category intelligence, not copying. The creative inspiration workflow shows how to turn that research into a repeatable creative brief.

The IAB's digital advertising standards publish benchmark CPMs and format performance by category — useful for sanity-checking your meta ads budget assumptions before you commit. If you are entering a category with €12 CPMs and budgeting for €4, the math won't work regardless of meta ads creative quality.

According to HubSpot's State of Marketing 2024, marketers who research competitor creative before launching meta ads campaigns report significantly higher first-campaign ROAS than those who build from scratch. The pattern is consistent: research first, spend second.

If you want programmatic access to competitor ad data (creative metadata, active ad counts, and platform coverage across Facebook, Instagram, TikTok, YouTube, and LinkedIn in one query), that's where the AdLibrary API earns its place. Meta's free Ad Library is adequate for one-off manual checks on a single brand. The moment you want structured data at scale or cross-platform coverage, you need something else. Meta's free API handles one platform. The moment you add TikTok, YouTube, or LinkedIn data into the same meta ads query, you need a paid solution. AdLibrary's Business tier starts at €329/month and includes API access with 1,000+ credits per month. See pricing for agency-scale research.

For structured competitor research workflows, the competitor ad research use case covers the full methodology. If you are running cross-platform meta ads research, the multi-platform coverage feature and cross-platform ad strategy use case are where to start. The unified ad search feature is the fastest path to cross-channel creative intelligence in one interface.

After Launch: Meta Ads Results in the First 30 Days

Your first meta ads campaign is a research exercise. You are buying data about your audience, your creative, and your offer. Once the learning phase exits and you have a clear creative winner, scale the budget 15–20% every 3–4 days, test one new creative per week, and build a retargeting layer off the warm audience you have accumulated. The cold audience ramp use case walks through the full 30-day meta ads protocol.

If your conversion rate is strong but CPA is still high, the bottleneck is upstream — your landing page or offer, not your meta ads. If CPA is improving but hook rate is low, you have a creative attention problem. These two diagnostics resolve 90% of first-campaign issues.

For a full diagnostic framework, the Meta ads not converting post covers every root cause. The Facebook ads for small business guide has a tighter version of this launch workflow for sub-€2,000/month budgets.

According to Forrester's B2C marketing benchmarks, brands that instrument conversion tracking before their first paid campaign cut their time-to-positive-ROAS by 40% on average. Set the Pixel up first. Launch the meta ads campaign. Read what it tells you. Iterate. That's the whole system for how to launch meta ads from scratch.