Every day you check your Meta Ads, you face the same questions: Should I increase budget on this campaign? Should I pause that one? Is this ROAS good enough? Should I wait longer before deciding?
Most advertisers don't have a system for answering these questions. They rely on gut feel, or they overthink every decision, or they make changes too fast and never let campaigns stabilize.
The Scale/Watch/Kill framework solves this. It's a simple decision system that categorizes every campaign, ad set, and ad into one of four buckets—each with a clear action attached.
No more analysis paralysis. No more wondering what to do. Just look at the verdict and act.
The Four Verdicts
Every piece of your advertising gets one of four verdicts based on your custom thresholds:
Scale
Above your target ROAS. These are winners. Consider increasing budget to capture more of this profitable traffic.
Watch
Profitable but not crushing it. Above breakeven but below your scale threshold. Monitor for improvement or optimize.
Kill
Below your minimum ROAS. Losing money. Pause immediately or make significant changes before continuing.
Learning
Not enough data yet. Below your minimum spend threshold. Let it run before making decisions.
That's the entire system. Four categories. Four actions. Everything else is just deciding where to set your thresholds.
Setting Your Thresholds
The framework only works if your thresholds reflect your actual business. Here's how to set them:
Minimum ROAS (the "Kill" line)
This is your breakeven point—the ROAS below which you're losing money on ad spend.
To calculate it:
- Take your average order value
- Subtract your cost of goods and fulfillment
- What's left is your contribution margin
- Minimum ROAS = AOV / Contribution Margin
Example Calculation
Average order: $80
Cost of goods + fulfillment: $30
Contribution margin: $50
Minimum ROAS = $80 / $50 = 1.6x
Anything below 1.6x ROAS and you're losing money. That's your Kill threshold.
Scale ROAS (the "Scale" line)
This is the ROAS at which you're happy to invest more. It should be meaningfully above breakeven—high enough that you're confident in the profitability even with some fluctuation.
Most businesses set this at 1.5-2x their minimum ROAS. If your minimum is 1.6x, your scale threshold might be 2.5x or 3x.
The exact number depends on your risk tolerance and growth goals. Aggressive growth? Set it closer to minimum. Conservative approach? Set it higher.
Learning Spend (the "Learning" threshold)
New campaigns and ad sets need data before you can judge them. The Learning threshold defines how much you let them spend before making decisions.
A good rule of thumb: set learning spend at 2-3x your target CPA. If you're aiming for $20 CPA, set learning threshold at $40-60.
This gives the algorithm enough data to optimize while limiting downside if the campaign isn't going to work.
How to Use the Framework Daily
Here's the workflow:
Step 1: Check your verdicts
Look at each campaign, ad set, and ad. What verdict does it have based on the data so far?
Step 2: Take the obvious actions
- Kill verdicts: Pause these immediately. Don't wait to see if they improve.
- Scale verdicts: Consider increasing budget by 20-30% if they've been consistently scaling for 3+ days.
- Learning verdicts: Leave alone. Let them spend to threshold before judging.
- Watch verdicts: Monitor but don't change. These need more time or optimization.
Step 3: Review Watch campaigns
Watch is the category that requires thought. These campaigns are making money but not enough to scale aggressively. Your options:
- Wait: Sometimes campaigns need time to optimize. Give them another few days.
- Optimize: Test new creative, adjust audience, or tweak the offer.
- Reduce: Lower budget to minimize exposure while keeping data flowing.
- Accept: Some campaigns never become "Scale" candidates. A steady 1.8x ROAS might be fine for brand awareness.
Why This Works Better Than Gut Feel
Most advertisers make one of two mistakes:
Mistake 1: Cutting too fast
You see a campaign with 0.8x ROAS after one day and pause it. But it only spent $30. That's not enough data. One purchase would have completely changed the numbers. You killed something that might have worked.
Mistake 2: Letting losers run too long
You see a campaign with 0.8x ROAS after two weeks and $500 in spend. "Maybe it'll turn around." It won't. You just burned $500 learning what should have been obvious after $100.
The framework prevents both mistakes. Learning threshold stops premature cuts. Kill verdict stops wishful thinking.
"The best advertisers aren't smarter than everyone else. They just have better systems for making decisions. This framework is one of those systems."
Applying It at Every Level
The framework works at campaign, ad set, and ad level—but with different implications:
At campaign level:
Verdicts tell you about overall campaign health. A Kill verdict here means the entire approach needs rethinking, not just tweaking.
At ad set level:
Verdicts tell you about audience and bidding. A Kill verdict might mean this audience doesn't convert, even if other ad sets in the same campaign are working.
At ad level:
Verdicts tell you about creative performance. A Kill verdict means this creative isn't resonating—try a different approach.
The key insight: you can have a Scale campaign with Kill ads inside it, or a Kill campaign with one Scale ad set. Looking at each level separately gives you more actionable information.
Common Questions
What if everything is in Learning?
You're probably not spending enough or being too patient. Either increase budgets to generate data faster, or accept that decisions will take longer at low spend levels.
What if everything is in Watch?
Your thresholds might be too aggressive. If 80% of campaigns never reach Scale, either you're running mediocre campaigns or your Scale threshold is unrealistically high. Try lowering it temporarily.
Should I ever ignore the verdict?
Rarely, but yes. Some situations warrant exceptions:
- New product launches might run at lower ROAS intentionally for market testing
- Brand awareness campaigns might have different success metrics
- Seasonal businesses might accept lower ROAS during off-peak times
But if you're constantly ignoring verdicts, you need different thresholds, not exceptions.
How often should I check?
Daily check-ins, but don't act every day. The algorithm needs stability. Check daily, act 2-3 times per week unless something is severely underperforming.
The Meta-Lesson
The real value of this framework isn't the specific verdicts—it's the discipline of having a system.
When you have clear rules for decisions, you:
- Spend less mental energy on each decision
- Make more consistent choices over time
- Avoid emotional reactions to daily fluctuations
- Learn faster because you can evaluate your rules, not just your campaigns
The framework gives you that system. Adapt it to your business, set your thresholds, and let the verdicts guide your decisions.
You'll make fewer mistakes, move faster, and probably sleep better knowing exactly what to do with every campaign.
Get automatic verdicts for every campaign
KillScale applies the Scale/Watch/Kill/Learn framework automatically. Set your thresholds once, and see instant verdicts for every campaign, ad set, and ad. No more spreadsheets. No more guessing.
Start Free — No Credit Card Required