Key takeaways
- A new pixel has no purchase history, so Meta's delivery has to learn from scratch. Your first month is about generating signal, not maximising ROAS.
- Get tracking, product pages and offer right before you spend. Ads can't fix a page that doesn't convert.
- Start with one Sales campaign, one broad ad set and several genuinely different creatives, plus a small fixed testing budget.
- Judge early weeks on leading signals (click-through, cost per add-to-cart, conversion rate) and later weeks on cost per purchase.
- Plan for six to twelve distinct creatives in the first month, because finding the first winning angle is the main job.
For a new brand with no pixel history, the first 30 days of Meta ads are about generating signal, not hitting a perfect ROAS. Set up tracking and a converting product page first. Then run one Sales campaign with a broad audience, a budget your margin can sustain for a month and six to twelve genuinely different creatives, judging early weeks on click-through, add-to-cart cost and conversion rate before cost per purchase.
Why is starting Meta ads from zero different?
Meta's delivery system learns who buys from the purchases it observes. An established store has thousands of past purchases, so its ads start from a strong base. A new store has none. Every early impression is partly exploration, which means:
- Costs are higher at first. Meta's help page on the learning phase says ad sets are less stable and usually have a higher CPA while learning.
- Results are noisy. With a handful of purchases, one lucky or unlucky day swings everything.
- You don't know your winning angle yet. Established brands know which message sells. You are still finding out.
That changes what success looks like in month one. The goal is to find the first ads and offer that sell profitably, and to build enough purchase data that Meta can find more buyers like them.
What do you need before you spend anything?
Every dollar spent before these are in place is partly wasted.
Tracking
- Meta Pixel and Conversions API connected through your platform's official integration, such as the Facebook & Instagram app on Shopify. See Facebook ads for Shopify stores and the Meta Pixel and Conversions API checklist.
- A test order that shows exactly one Purchase in Events Manager, with the right value and currency.
- Domain verification in your business portfolio.
A page that can convert
- Real photos, a clear headline about what the product does, price, delivery times and costs, and a returns policy.
- Some social proof, even early: a few reviews from friends-and-family buyers or samples, a founder story or press if you have it.
- Fast on a phone. Most Meta traffic is mobile.
Numbers
- Break-even CPA = average order value × gross margin % − shipping, fees and packaging per order. The break-even ROAS calculator works it out.
- A launch budget you can sustain for 30 days without panicking if week one is bad.
An offer worth clicking
New brands have no reputation to lean on, so the offer does more work. A first-order discount, free shipping, a bundle or a risk reducer such as an easy returns promise gives cold audiences a reason to try you. Make sure your margin can carry it.
How much should a new brand spend in the first month?
Spend what your margin and cash can sustain for 30 days. Starting and stopping wastes the learning you have paid for.
A simple way to set it:
- Find your target CPA, usually below break-even, for example $25 if break-even is $33.
- Decide how many purchases you want in month one to learn from. Even 30 to 60 purchases tell you a lot about angles and audiences.
- Monthly budget ≈ target purchases × expected early CPA. Expect early CPAs above target. If you plan for 40 purchases at an early CPA of $35, that is about $1,400 for the month, or roughly $47 a day.
Meta's guidance is that an ad set usually exits learning after about 50 results in the week after its last significant edit, which at a $25 CPA means about $179 a day. Most new brands can't fund that from day one, and they don't need to. Below it, the ad set will show Learning limited and results will be noisier, so you judge over longer windows. Our full method is in how much to spend on Facebook ads, and the Meta ads budget calculator runs the numbers.
How should a new brand structure its first campaign?
Simple, with all the learning in one place:
| Setting | Week 1 setup |
|---|---|
| Objective | Sales, optimising for Purchase |
| Campaigns | One main campaign, plus a small testing campaign from week 2 |
| Ad sets | One broad ad set |
| Audience | Advantage+ audience, or your countries and adult ages only, with an optional interest suggestion |
| Placements | Advantage+ placements |
| Bid strategy | Highest volume, no cost cap yet |
| Ads | Four to six genuinely different concepts |
Why broad on a new pixel? Because narrow audiences cost more per impression and give a new account even less room to learn, and because your creative attracts the right people. A single relevant interest used as a suggestion can give delivery a starting point. Our broad vs interest vs lookalike targeting guide explains how suggestions work under Advantage+ audience.
Why not an Advantage+ sales campaign on day one? You can use one, since in current Ads Manager a Sales campaign with Advantage+ on for budget, audience and placements is one. But a brand-new pixel gives it little to learn from, and it is designed for scaling proven ads. Our Advantage+ sales campaigns guide covers when it fits.
How many creatives does a new brand need?
More than you think. Your biggest unknown is the angle, the reason someone buys, and you only find it by testing several.
Aim for six to twelve distinct concepts in the first 30 days, spread across:
- Angles: the problem it solves, the result or transformation, what it is made of, how it compares with alternatives, who made it and why, what customers say.
- Formats: a short vertical video (9:16), a static image (4:5), a UGC-style clip of someone using it and, for several products, a carousel. The Meta ad sizes guide lists the specs.
- Hooks: the first line of copy and the first two seconds of video. Our facebook ad copy formulas give you starting points.
"Distinct" matters. Five versions of the same video with different captions are one concept. You want ads that would make different people stop scrolling.
Where do new brands get ideas? Study competitors' long-running ads in the Meta Ad Library. Ads that have run for months are usually profitable, which tells you which angles work in your category.
The 30-day plan, week by week
Week 1: launch and watch the leading signals
- Launch the main campaign with four to six concepts.
- Don't edit for at least three days. Every significant edit restarts learning.
- Watch leading signals rather than ROAS: CTR, cost per add-to-cart, landing page views and conversion rate on the site.
- Fix anything broken: a page error, a checkout problem, an event not firing.
Judge on: are people clicking and adding to cart at a reasonable cost? Purchases are a bonus this week.
Week 2: cut clear losers, start testing
- Pause concepts that have spent two to three times your target CPA with no purchase and poor click-through.
- Start a small testing campaign on ad set budgets, about 20 to 30% of spend, with two or three new concepts built on whatever is showing promise.
- Keep the main ad set's budget steady.
Judge on: cost per add-to-cart and the first purchases. Which angle is pulling ahead?
Week 3: double down on the angle
- Move test winners into the main campaign in one batch, not one ad at a time.
- Make new variations of the best angle: new hooks, new first frames, a different person on screen.
- If cost per purchase is near or below target for several days, raise the main budget by around 20%.
Judge on: cost per purchase over the last seven days, and conversion rate on the site.
Week 4: decide what scales
- Review the whole month: total spend, purchases, cost per purchase, average order value and new-customer revenue in your store.
- Keep winners, stop losers and plan next month's tests around what worked.
- If one ad set is profitable and stable, plan to scale in steps of 20 to 25% every two to three days. Our guide to scaling Meta ads without killing ROAS continues from here.
Judge on: is there at least one angle and one offer that sell at or near target CPA? If yes, month one worked.
What should you judge a new brand's ads on?
Pick metrics that match the amount of data you have:
| When | Judge on | Why |
|---|---|---|
| Days 1–7 | CTR, cost per add-to-cart, landing page views, site conversion rate | Purchases are too few to read |
| Days 8–14 | Cost per add-to-cart and first cost per purchase, by concept | Angles start to separate |
| Days 15–30 | Cost per purchase against target, by concept and overall | Enough purchases to compare |
| Month end | Store revenue vs total ad spend (MER), new customers | Whether the business grew |
Keep two guard rails:
- A spend-based rule for every test, such as two to three times target CPA, so you don't kill ads on one bad day or keep losers forever.
- A monthly loss you can accept. Month one rarely hits target ROAS. Decide in advance how much you are prepared to spend to learn.
What are the most common new-brand mistakes?
- Spending before tracking works. Meta can't learn from purchases it can't see.
- Too many ad sets. Splitting $50 a day across five audiences means none of them learns.
- Too few creatives. Two ads isn't a test, it's a coin toss.
- Judging on day two. Early results swing wildly in both directions.
- Editing constantly. Every significant edit restarts learning.
- Boosting posts instead of Sales campaigns. Boosts usually optimise for engagement, not purchases.
- Narrow interest targeting. Expensive reach and fragmented learning on an account that has little data to spare.
- No offer. Cold audiences need a reason to try an unknown brand.
- Stopping after a bad week. Restarting later means learning from scratch again.
New brand launch checklist
- Pixel and Conversions API connected, test order shows one Purchase
- Domain verified
- Product page with photos, price, delivery, returns and some proof
- Break-even and target CPA calculated
- 30-day budget set and committed
- One Sales campaign, one broad ad set, highest volume
- Four to six distinct concepts at launch, six to twelve over the month
- Spend-based judging rule written down
- Weekly review date in the calendar
Getting creative volume without a studio
The hardest part for most new brands is making enough distinct creative. Fly ROAS reads your website to learn your brand, studies long-running competitor ads in the Meta Ad Library and builds image and video ads from your real product photos, then can launch them to your own ad account, paused for your approval if you prefer.
Buy data first, then buy growth
A new brand's first month of Meta ads is an investment in information: which angle sells, which offer converts and who buys. Get tracking and the product page right, keep the structure simple, test enough distinct creatives, judge on signals that match your data and commit to the full 30 days. By the end you should have at least one ad that sells near your target, and an account that knows who to show it to.
Frequently asked questions
Can you run Meta ads with no pixel data?
Yes. Meta's delivery system learns as purchases arrive, so a new pixel simply starts with less to go on. Expect higher costs and noisier results at first, keep the structure simple with one broad ad set and judge early weeks on click-through, add-to-cart cost and conversion rate.
How much should a new brand spend on Facebook ads in the first month?
Spend what your margin and cash can sustain for the full 30 days. A simple method is to decide how many purchases you want to learn from, often 30 to 60, and multiply by an early CPA somewhat above your target. Committing to the month matters more than the exact figure.
Should a new brand use interest targeting?
Broad targeting or Advantage+ audience usually works better, because narrow audiences cost more and fragment learning. A single relevant interest used as a suggestion can give a new pixel a starting point, and you can test removing it once purchases flow.
How many ads should a new brand test?
Plan on six to twelve genuinely different concepts in the first 30 days, launching four to six at the start. Vary the angle, format and hook, because finding the reason people buy is the main job of the first month.
Should I optimise for add to cart on a new pixel?
Only if purchases are very unlikely at your budget. Add-to-cart optimisation exits learning faster but learns to find people who add to cart, who are not always buyers. If you can afford roughly one target CPA a day, start by optimising for Purchase.
