Step 1
What to test next on $100 a day.
Everyone says "it's your creative" and "you need 30 new ads". At $100 a day, 30 is a lot. Here's how many ads to run, how long to wait, when to cut and when to scale, with a calculator that does the math.
- 1Your numbers
- 2The 4 calls
- 3What to test
- 4Your 14 days
Step 2
The 4 calls, and nothing else
Every new ad ends in one of these. Check once a day, decide once a week.
Wait
It has spent under $30, or had under 1,000 impressions, or it's been live under 7 days.
Do nothing. Don't pause it, don't edit it. Early numbers swing a lot.
Cut
It spent your cut line from step 1 with no sale, or its cost per sale is more than 2x your target.
Turn it off. Write down what you tested so you don't test it again.
Too close
Its cost per sale is within about 30% of your best ad.
Not a winner yet. Leave both running 3 to 5 more days, then look again.
Scale
It beats your best ad's cost per sale by more than 30%, after $30 and 7 days.
Raise its budget by 20%. Do it again every 3 to 5 days while it holds.
Step 3
Test in this order
- 1. The first 2 seconds. Same ad, new opening. It's the cheapest test and moves results the most.
- 2. The format. Talking to camera vs hands only vs before and after. Same message, new way of showing it.
- 3. The angle. A new reason to buy: a different problem, a different person, a different moment.
- 4. The offer. Bundle, free gift, first order discount. Test last, it costs you margin.
One change per test. If you change the opening and the offer at once and it wins, you don't know which one did it.
Step 4
Your 14 days, ready for Google Sheets
Built from your numbers above. Copy it, open a new Google Sheet, click cell A1 and paste.
Bonus: examples
12 tests ranked for a $100 a day brand
What to change, what to keep the same, and what each test tells you, in the order to run them. Plus the weekly review.
Want these made for your brand instead? Get 2 free storyboards or have us make all 15 ads.
Quick answers
Why wait $30 per ad?
Below that, one lucky or unlucky sale swings the numbers too much. $30 and 1,000 impressions is the least we judge an ad on. If your product sells above $60, wait longer: use 1.5x your target cost per sale.
Why not test 30 ads at once?
At $100 a day, 30 ads get about $1 each. None of them spends enough to tell you anything for a month. Fewer ads, judged properly, beats more ads judged on noise.
Should I test in a new campaign?
Keep it simple: one test campaign next to the one that's working. Move a winner into your main campaign once it passes the Scale call.
What counts as a sale?
Whatever you optimize for: a purchase on most stores. Use the cost per purchase from Ads Manager, not a guess.
Next step
Get 2 free storyboards for your store
Send your store link. We pick your best-selling product, look at what's running longest in your category, and send you 2 storyboards for it: every shot, every line, ready to film or hand to a creator.
Rather not make them yourself?
We make these for your product
Everything on this page, done for you: we find the ads that last in your category and rebuild them for your product.
- 9 video ads and 6 static ads, ready to test
- A test plan: which 3 to launch first and when to cut
- If none of the 9 videos beats your best ad within 14 days, you get 3 more hooks free.
$450
Want to see the quality first? See what we make
More free guides
- Remake your rival's longest-running adFind the ad a rival has paid for all year, take it apart, rebuild it for your product.
- Turn your product page into a 15-second ad scriptPaste your page, get a script, a shot list and 3 openings to test.
- The 565-day ad, broken downA body care ad that has run since April 2025. Every shot, why it lasts, and how to remake it.
- The creator brief that gets the right take the first timeYour AI interviews you and writes the 1-page brief and kickoff message. Plus a review checklist.
