By now, you have your numbers, your audience library, and two cold ads — one outcome-led and one objection-led. The next step is to put them into a test campaign. This is the first time your ads will face a real audience, and the way you structure this launch will decide whether you find a winner or waste money.
Most advertisers go wrong here. They mix different audiences into one ad set, they run too many creatives, or they spend so little that Meta never has the data to optimise. The result is chaos: one day looks great, the next looks dead, and they have no idea why. You avoid that by testing with discipline.
The campaign structure
Your first cold test should be a single Sales campaign with ad set budgets (ABO). Start with three ad sets, each targeting a different cold audience type. For example: one competitor cluster, one lookalike, one interest stack. Keep them separate so you can see which type of source is pulling buyers.
Inside each ad set, you put the exact same two ads: the outcome-led creative and the objection-led creative. Every audience sees both ads. That way, if one ad performs much better, you know it’s the creative. If one audience performs better, you know it’s the targeting.
Do not stack competitor, lookalike, and interest together in the same ad set. If you do, you’ll never know which one actually worked.
Budgets
Budget matters. If your product is £50, spending £5 a day won’t tell you anything. Meta needs enough budget to actually get close to a purchase. A safe starting point is £10–£15 per ad set per day. That means three ad sets will cost £30–£45 per day. It’s enough to see data without committing thousands.
Think of it this way: each ad set should be allowed to spend at least the price of your product or service before you judge it. If your consultation is £250, you won’t know anything at £20 spent. If your product is £49, spending £49 with no sales tells you something useful. This is why knowing your CPA ceiling from Lesson 1 matters — you already know exactly how much you can afford to spend before pulling the plug.
Rules for killing and keeping
Don’t rely on hunches. Set your rules now. If an ad set spends the full product price without a sale, you pause it. If CTR(all) stays under 2% once half the product price has been spent, you pause it. If you’re hitting or beating your max CPA, you keep it running. If you’re not at CPA yet but CTR is strong and add-to-carts are coming in, you let it run. These rules remove emotion. The numbers decide, not your mood.
What to expect
Don’t expect every audience to work. In fact, most won’t. In almost every account, one or two cold audiences will carry the rest. That’s why you only test three at a time. You’re looking for the first sign of life — one audience and one creative that starts pulling sales at or under your CPA target. That’s your cold winner.
Also don’t expect instant results on day one. Sometimes it takes a few days of spend for Meta to find traction. That’s why you watch the rules, not the daily swings.
Task
Set up a Sales campaign in Ads Manager. Create three ad sets, each with a different cold audience from your library. Put the same two cold ads inside each ad set. Start with a budget of £10–£15 per ad set per day. Apply your rules: pause if it spends the product price with no sales, or if CTR stays below 2%. Keep if CPA is at or below target, or if strong engagement signals are showing.
This is your first cold testing campaign. From here, you’ll learn which audiences and which ads are worth scaling.
