A 'conversion event' is just Meta's name for 'the moment someone becomes valuable to you' — a form submit, a purchase, a booking. The Pixel is a small snippet on your website that watches for that moment and reports it back to Meta. The Sales objective you'll pick optimizes toward this 'Lead' event, so if it doesn't exist or doesn't fire, Meta is optimizing toward nothing and the budget is wasted. This is the single most-skipped step by beginners — nail it and you're ahead of most people running ads.

The Event Setup Tool is Meta's no-code way to attach a conversion event to a click or a page view. It's perfect for someone who 'never ran ads' and doesn't want to touch code. Behind the scenes it's doing the same thing a developer would hard-code, just visually. If your site is on Shopify or another platform with a native Meta integration, the Lead event may already exist — in that case this slide is just a confirmation, not new work.

Verifying beats assuming. Test Events gives you live confirmation that when a real person does the thing, Meta receives the 'Lead' signal — which is exactly what the algorithm needs to find more people like them. Common gotchas it catches: the event mapped to the wrong button, the event firing on every page (not just the conversion), or the Pixel not loading at all. Catch it now, not after wasting $50/day.

Meta's AI only gets smart when you feed it a clear 'win' signal. That win is the Lead event. The Pixel is the little piece of code on your site that watches for it and reports back 'hey, that click converted.' Without it, you're paying for clicks and hoping — with it, Meta actively hunts for more people like the ones who converted. This single step is the difference between a campaign that improves over time and one that just burns money.

Names are how you read reports at a glance without opening anything. The pattern is Brand_Funnel_Strategy_Budget_Date at the campaign level, then Brand_Funnel_Audience_Geo_Age_Event at the ad set, and Brand__AD##__label for each ad. When you scale to several campaigns you'll thank yourself — you'll instantly see which creative and which audience won straight from the names.

The objective tells Meta what a 'win' looks like. Sales (optimizing for the Lead event on your site) points the algorithm at people likely to actually convert — real intent — not just cheap clicks. The Leads objective is for on-Facebook instant forms; since we're sending people to your own website with a Pixel, Sales is the correct, higher-intent choice.

Doing this now (instead of leaving the default 'New Sales Campaign') means you never end up with a graveyard of 'Copy of Copy of New Campaign'. The date stamp (2026-07) lets you compare this month's test against next month's cleanly.

CBO means you hold the budget at the campaign level and let Meta pour it into whichever ad is performing best in real time, instead of you guessing. $50/day is a solid starting point for a conversion campaign: enough daily conversions to escape the 'learning phase' within roughly a week, but not so much that you overspend before you have data. Bigger isn't better early — data is.

This is where you tell Meta exactly what to optimize for. 'Maximize conversions' + the Lead event = 'go find me the most conversions you can for $50/day.' Because the budget is already set at the campaign level, you won't see a budget field here — that's correct, not a bug.

In 2026, with strong creative, broad beats hand-picked interests almost every time. Your ad copy self-selects the right person better than any interest checkbox. Advantage+ then uses your conversion event to learn who actually converts and floods delivery toward them. The controls (location, age) are just guardrails so you don't pay to reach the wrong people — not targeting.

This is exactly why creatives come in different shapes (square/portrait for Feed, 9x16 for Reels/Stories). Advantage+ placements let Meta drop the right aspect ratio into the right spot and find the cheapest conversions across all of them. Restricting placements manually usually just raises your cost per result.

One clean ad set for a single broad audience is the right call here. You're not splitting audiences yet — you're testing which CREATIVE wins first. The name encodes the audience (ADV-BROAD), geo, age, and optimization event so it reads like a sentence.

The identity is the face of the ad — the Page/IG account it appears to come from. Using your real accounts builds the page's social proof (likes, comments, shares accumulate on the real brand). 'Manual upload / Single image' keeps things simple and predictable. You build ONE ad completely, then duplicate — so you only fill the shared settings once.

3 ads is a healthy starting spread: enough variety for Meta to find a winner, few enough that each still gets real spend inside a $50/day budget (spread across 10+ ads, none get a fair test). We keep the COPY identical across all of them and only change the CREATIVE — because changing one thing at a time is the only way to know what actually won. Once one wins, that's when you test new angles with that winner.

This is your master ad — build it carefully once, because the other 2 are just duplicates with the image, name, and tracking tag swapped. Upload the feed image as the main asset, then use 'Edit per placement' to drop the 9x16 into Stories/Reels so the vertical spots get a full-screen creative instead of a cropped one. The utm_content (ad01_variation-a) is what lets you later say 'Variation A produced X results' in your own analytics.

Duplicating keeps every ad identical except the one thing you're testing (the creative), which is what makes the test fair — a clean A/B/C. You only touch the name (so reports read clean), the image (the actual variable), and utm_content (so this ad's results are traceable to 'Variation B'). Always duplicate from the SAME master (AD01), never from a previous duplicate, so a stray edit can't quietly carry forward.

Duplicating keeps every ad identical except the one thing you're testing (the creative), which is what makes the test fair — a clean A/B/C. You only touch the name (so reports read clean), the image (the actual variable), and utm_content (so this ad's results are traceable to 'Variation C'). Always duplicate from the SAME master (AD01), never from a previous duplicate, so a stray edit can't quietly carry forward.

Advantage+ creative enhancements can nudge performance, but the aggressive ones (auto-added text, music, cropping) will wreck a carefully designed creative — cropping can chop your headline, and auto-text fights your baked-in copy. You spent real effort on these designs; protect them. Once you're comfortable, you can test enhancements ON as its own experiment later.

Getting flagged once is normal, not a crisis. The review bot is strict, especially about claims. If rejected, the fix is almost always in the wording: remove anything that sounds like a guarantee or a sensitive claim and resubmit. Keep the language clear and honest.

When a campaign launches, Meta is actively experimenting to figure out who converts — that's the learning phase, and results are volatile and expensive during it. Every time you edit something important, you reset that clock and throw away what it learned. The hardest, most valuable ad-management skill is patience: build it right, then let the machine cook.

Read them top-down like a diagnosis. Cost per result tells you IF it's working; the others tell you WHY. High cost + low CTR = it's the creative (test a new hook). High cost + good CTR but few conversions = it's the landing page (the ad did its job, the page didn't). Separating 'ad problem' from 'page problem' is the skill that makes you dangerous — most people blame the ad when the page is the leak.

Rules beat feelings. The reason most people lose money on ads isn't strategy — it's fiddling: killing winners too early, scaling too fast, changing three things at once so they can't tell what worked. Decide the rules while you're calm (now), then just execute them. When a creative wins, that's when you test a new angle on that winner — that's your next campaign.