Most first campaigns fail in week one, and almost always for the same reason: the person running them reacts to noise. The platform needs time and stability to learn, and the instinct to fix things is exactly what prevents it. Here is a plan that accounts for that — including the specific moments where doing nothing is the correct action.
Before you launch: the half hour that decides everything
Write the description honestly
Two sentences about who your customer is and why they choose you. This improves the campaign more than any tool choice. Generic input produces generic output, every time.
Work out customer value
What a customer is worth to you, and roughly what fraction of enquiries become customers. Without these two numbers you cannot tell a good result from a bad one — and no tool can supply them.
Verify the conversion event
Open the exact page your ad will point at, with the pixel debugger running, and confirm the event fires. Five minutes, and it is the most common silent failure in advertising.
Check the budget against the floor
Platform minimums apply per ad set and are roughly three times higher for conversion objectives. Below the floor the campaign is refused with an error that never explains why.
Week one: launch, then leave it alone
One campaign. One objective — the one matching what you actually want, not the one with impressive-looking reach. One properly funded ad set. Launch it and then, genuinely, stop touching it.
Hours 0–5: nothing happens
A new click or conversion ad set typically takes four to five hours before its first impression. This is normal. Reach objectives start in minutes, which is why comparing the two makes yours look broken.
Day 1: the graph shows zeros
A flat zero delivery curve and an estimated audience of zero on a brand-new ad set means nothing at all. Healthy campaigns show the identical signature. We used this twice as evidence of a problem and were wrong both times.
Days 2–7: numbers, but not meaning
You will have data. It will not yet describe your campaign — the platform is still learning who to show it to. Early cost per result is not your cost per result.
💡 The single most valuable rule of week one: do not edit the ad set. Every edit restarts the platform's evaluation. We watched an ad set edited four times in quick succession serve zero impressions for fourteen hours while the same account delivered thousands elsewhere — everything read healthy throughout. A fresh ad set, created once and left alone, delivered within the hour.
The only week-one interventions that are justified: fixing a rejected ad, fixing billing, and fixing a conversion event you discover is not firing. Everything else waits.
Week two: read, do not react
Now the numbers start to mean something. Look at three things and nothing else:
- 1Cost per result — not cost per click. Hold it against what a customer is worth to you. This is the only judgement that matters.
- 2The gap between clicks and landing page views. If a large share never arrive, you are buying low-quality attention. On one campaign we saw 45 clicks produce 28 arrivals — nearly 40% lost before the page rendered.
- 3Whether anything downstream moved. Form starts, calls, messages, add-to-carts. If nothing after the visit happened, the click cost is irrelevant.
Resist restructuring. If something is genuinely wrong — a broken destination, an ad rejected, a budget below the floor — fix it. Otherwise let it run the full fortnight.
Week three: add a comparison
Now you can learn something. Add a second audience segment, or a second keyword theme, in its own ad set — not by editing the first one. You are now able to answer the actual question: which customer is cheaper for me to reach?
Budget for it properly. Platform minimums apply per ad set, so two segments means two funded ad sets. If your budget cannot carry both, keep running one — two starved ad sets teach you less than one healthy one, and the temptation to split is the most common self-inflicted wound at small budgets.
And if the first campaign is genuinely not working after two full weeks of being left alone, this is the point to rebuild rather than repair. Duplicate, change what needs changing, and let the new one run clean. Do not keep editing the old one.
Week four: decide something
Calculate cost per customer
Cost per lead divided by your close rate. That is the number to hold against customer value — not cost per lead, which tells you what you paid rather than what you got.
Compare your two segments
One will be cheaper. Move budget toward it, by adjusting budgets rather than rebuilding audiences.
Decide about the second channel
If search is producing and you have not tried social, or the reverse, week four is when you have enough baseline to judge whether the second channel beats the first.
Or stop
A real option. If cost per customer exceeds customer value after a month of properly run campaigns, advertising at this budget is not your growth channel right now. Knowing that in thirty days is worth the thirty days.
The mistakes that account for most first-month failures
- 1Editing the ad set repeatedly in week one. By far the most common, and it looks like diligence.
- 2Splitting a small budget across several ad sets, none of which reach the minimum properly.
- 3Judging on cost per click instead of cost per customer.
- 4Launching with a conversion event that does not fire, so the platform optimises toward nothing.
- 5Picking an awareness objective because the reach numbers look impressive. On one account, awareness cost five times more per click than traffic in the same week.
Built once, validated, then left alone
Adyft builds the campaign, checks the budget against the platform floor, validates the payload before launch, and does not rewrite live ad sets. 14-day free trial, no card.
Start free — no card needed