When I launch a new product, I treat the first few days like a street fight. No rules committee, no fancy dashboards, no "let it optimize for a week." Just instinct, data, and speed. Every time I get three sales, I bump the budget. No emotions. Just math.
This is the exact system that took one of my stores from zero to $50K in 8 days, tripled a monthly challenge target to $170K, and eventually produced $50K and $82K single days on TikTok ads alone. Nothing here is theory. Every number is one I actually spend against.
Step 0: The launch structure
On a fresh ad account, you do not know the account's DNA yet, so you run all three campaign types at once and let the account tell you what it likes:
- 1 ABO at $21.77/day. One ad group per creative, one video per ad group. Every creative gets its own budget lane, so a strong hook cannot be starved by a weak one.
- 1 CBO at $51.77/day. Five ad groups, all creatives inside each. The algorithm allocates, you observe.
- 1 SPC at $51.77/day. One bucket, minimum 6 videos, fully hands-off. TikTok's automation does everything.
On an aged account where you already know which structure wins, skip discovery: run 3 campaigns of the winning structure only, each at $51.77.
The 5-5-5 formula
Most people either kill ads too early or bleed money hoping. The 5-5-5 formula removes the guessing. You evaluate in $5 increments, and each $5 has to earn the next one.
| Checkpoint | What must be true (US market) | If yes | If no |
|---|---|---|---|
| First $5 spent | CPC under $1.00, CTR over 1%, CPM under $8 to $10 | Spend the next $5 | Kill or fix the creative |
| Next $5 | At least 1 add to cart | Spend the next $5 | Kill, or diagnose the page |
| Next $5 | A sale | Feed it, you have a candidate | Apply the kill rules below |
EU and UK markets run meaningfully cheaper on every one of these numbers. Adjust down, not up.
The kill rules
Before you spend a dollar, know your break-even cost per acquisition:
BE-CPA = selling price minus product cost minus payment fees. Example: $43 price, $18 COGS, $2 fees = $23 BE-CPA. That is the line between scaling and donating money to TikTok.
- Standard kill: no profit by $12 to $13 spend, kill it.
- SPC gets more grace: automation needs room, kill at $15 to $18 if there is no turnaround.
- The wait-and-see zone: no sale by $15 but the metrics are green and carts are coming? You can push to around $22 total. That is the ceiling.
- The real rule: move on fast, but never blindly. Every kill gets ten seconds of diagnosis: was it the creative, the offer, or the page? That answer is worth more than the $13 you just spent.
The 4 days
Day 1: feed the winners
Every 3 sales, bump the budget. CPA below break-even means add fuel. This is mechanical. The moment you start negotiating with a losing ad ("maybe it just needs time"), you have already lost.
Day 2: duplicate before breakfast
Let everything run overnight. Late-night TikTok is full of impulsive buyers, and the data they leave behind matures your signal. In the morning, duplicate your best campaign 10 to 20 times, set them to start at 8am, and kill anything that has not produced profit by $12 to $13.
Day 3: the pixel knows
By now the pixel understands who buys. Duplicate the winning campaign 30 to 40 times with budgets between $101.77 and $133.77. Watch it closely: TikTok spends fast, and a mature pixel with real budget behind it is either a money printer or a money shredder. The dashboard tells you which within hours.
Day 4: repeat without getting clever
Bump what works, trim what does not, watch CPAs against your break-even. Discipline beats creativity here. The system already found the winner; your only job is not to break it.
Account DNA: why good products die on bad accounts
Every ad account has its own DNA. Some accounts print for fashion and choke on gadgets. So never test the same product on the same account twice and call it dead. Rotate:
- Product 1 on accounts A and B
- Product 2 on accounts C and D
- Product 3 on accounts E and F
- Product 4 mixed, say B and D
Keep notes on which accounts perform for which niches. Poor results do not always mean a bad product. Sometimes it is just the wrong account. The full system is in the multi-account testing playbook.
Diagnosis table: read the metrics like a chart
| Symptom | Real cause | Fix |
|---|---|---|
| High CPC, low CTR | Weak offer. No urgency, no reason to stop scrolling | Rebuild the offer, adjust price, check what abandoned checkouts say about shipping cost |
| High CPM | Ad reads as an ad, not native content. If it is high across ALL ads, the account itself may be flagged from negative feedback | Make creatives organic. Account-wide? Move to a fresh account |
| Great CPC, CTR, CPM but no carts or sales | The website is the leak, not the ad | Fix the mismatch between ad and page, remove fake timers, add a real size chart, real reviews, lifestyle images, clear shipping and refund policies, cut pre-checkout friction |
That third row is the one nobody wants to hear. If the ad metrics are green and the money is not arriving, stop touching the ads. Read the pricing playbook next, because price is usually the first thing broken on the page.