Kill a Facebook ad when link CTR drops below 1%, CPA exceeds 2x your target, or frequency climbs past 3-4 with declining engagement. Those are the three fastest kill signals. But there are five metrics total that should drive every cut decision, and most advertisers either ignore them or misread them — holding onto losers for days or weeks while their margin evaporates.
This guide covers the five metrics that matter, the exact thresholds for each one, how long to wait before pulling the trigger, and what to do after you kill an ad so your next one performs better. No gut feelings. Just data.
The 5 Kill Metrics: What to Watch and When to Cut
Every ad in your account should be evaluated against these five metrics. Not one of them. Not whichever one you happen to check. All five, together. Here is the full framework with the thresholds that separate a struggling ad from a dead one.
1. Click-Through Rate (CTR)
Kill signal: link CTR below 1% after 1,000+ impressions. Link CTR measures the percentage of people who saw your ad and actually clicked through to your landing page. It is the most direct measure of whether your creative and copy are compelling enough to earn attention.
A link CTR below 1% on cold traffic means the ad is not resonating. The hook is weak, the image or video is not stopping the scroll, or the offer is not interesting enough to the audience. No amount of budget will fix a fundamentally uncompelling ad — it will just burn money faster.
Important: use link CTR, not all-CTR. Facebook's default “CTR (all)” metric includes reactions, comments, shares, and video views. That number is inflated and misleading. Link CTR only counts clicks to your destination — the clicks that can actually become conversions. You can find this metric in your Ads Manager dashboard by customizing your columns.
| Link CTR Range | Signal | Action |
|---|---|---|
| Below 0.5% | Dead on arrival | Kill immediately |
| 0.5% - 1.0% | Underperforming | Kill after 2,000 impressions if no improvement |
| 1.0% - 2.0% | Acceptable | Monitor other metrics |
| Above 2.0% | Strong | Let it run; focus on CPA and ROAS |
2. Cost Per Acquisition (CPA)
Kill signal: CPA exceeds 2x your target for 5+ consecutive days. Your target CPA should be derived from your unit economics, not from an arbitrary number. If your average CPA makes each sale unprofitable after accounting for COGS, shipping, and ad spend, the ad is losing you money on every conversion.
One or two days of high CPA can be noise — especially if you recently changed your budget or creative. But five consecutive days of CPA at double your target is a pattern. The algorithm has optimized for the audience it can reach at that price, and it is not going to magically get cheaper.
The 2x rule is a maximum, not a target. If your target CPA is $30 and you are consistently at $45, that is already a problem worth addressing — even though it has not hit the 2x threshold. The 2x rule ($60 in this case) is the point where you stop testing and start cutting.
3. Frequency
Kill signal: frequency above 3-4 with declining CTR and rising CPA. Frequency measures how many times the average person in your audience has seen your ad. At low frequencies (1-2), people are seeing fresh content. Above 3-4, they have seen the same ad multiple times and start ignoring it — or worse, hiding it.
High frequency by itself is not always a death sentence. Retargeting campaigns can run profitably at frequencies of 5-8 because the audience is small and highly qualified. But for prospecting campaigns targeting cold audiences, frequency above 3-4 almost always correlates with rising CPA and falling CTR. That combination — high frequency plus declining performance — is textbook ad fatigue.
The fix for high frequency is new creative, not more budget. Increasing spend on a fatigued ad just accelerates the fatigue. If your frequency is climbing, queue up new creatives before the performance collapse hits.
4. Return on Ad Spend (ROAS)
Kill signal: ROAS below your break-even threshold for 7+ days. Your break-even ROAS depends on your margins. If you have 50% gross margins, you break even at 2:1 ROAS. At 25% margins, you need 4:1 just to cover costs. Every dollar spent below break-even is a dollar lost.
ROAS is the most definitive kill metric because it directly measures profitability. CTR, CPA, and frequency are leading indicators — they warn you that something is wrong. ROAS is the lagging confirmation: the ad is not making money.
| Gross Margin | Break-Even ROAS | Kill If Below For 7+ Days |
|---|---|---|
| 20% | 5:1 | Below 5:1 |
| 25% | 4:1 | Below 4:1 |
| 33% | 3:1 | Below 3:1 |
| 50% | 2:1 | Below 2:1 |
| 66% | 1.5:1 | Below 1.5:1 |
If you do not know your break-even ROAS, stop reading and calculate it now. Without that number, every kill decision is a guess. True Margin's free ROAS calculator can give you the exact figure in under a minute.
5. Ad Relevance Diagnostics (Quality, Engagement, Conversion)
Kill signal: “Below Average” on two or more of the three relevance diagnostics. Facebook replaced the single relevance score with three separate diagnostics: Quality Ranking, Engagement Rate Ranking, and Conversion Rate Ranking. Each one compares your ad to competitors targeting the same audience.
A single “Below Average” ranking might be fixable with a creative tweak. But two or three “Below Average” rankings mean the ad is fundamentally mismatched with the audience. Facebook is telling you that competitors are outperforming you on quality, engagement, and conversions — and the algorithm will charge you more to compensate. That means higher CPMs, higher CPAs, and lower ROAS. Kill it and rethink the creative or audience.
How Long to Wait Before Killing (The Learning Phase)
Never kill an ad during Facebook's learning phase unless it has spent 2-3x your target CPA with zero conversions. The learning phase is the initial period where Facebook's algorithm is figuring out who to show your ad to. Performance during this phase is unreliable — CPAs swing wildly, delivery is inconsistent, and ROAS can look terrible even for ads that will become winners.
The learning phase typically ends after the ad set accumulates roughly 50 conversion events in a 7-day window. For most ecommerce advertisers, that means 5-10 days depending on daily budget and conversion rate.
Killing during the learning phase is the single most common mistake advertisers make. They launch an ad, see a $75 CPA on day two (when their target is $30), panic, and turn it off. But that $75 CPA was based on 3 conversions — nowhere near enough data to judge. The same ad might have settled at $28 by day seven.
The exception: if an ad set has spent 2-3x your target CPA without generating a single conversion, you do not need to wait for the learning phase to complete. Zero conversions at that spend level is a clear enough signal. An ad set that spent $90 with a $30 CPA target and has zero conversions is not going to turn around.
| Scenario | Wait or Kill? | Why |
|---|---|---|
| Day 2, high CPA, 3 conversions | Wait | Still in learning phase — not enough data |
| Day 3, spent 3x CPA target, 0 conversions | Kill | Zero conversions at high spend = wrong audience or creative |
| Day 7, 50+ conversions, CPA 2x target | Kill | Learning phase is over and CPA is too high |
| Day 5, 25 conversions, CPA slightly above target | Wait | Still building data — give it 2-3 more days |
When to Let a New Campaign Run
New campaigns deserve patience — but not blind faith. The key is knowing the difference between normal early-stage volatility and genuinely bad performance.
Give new campaigns a full 7 days before making any kill decisions. During this window, focus on leading indicators (CTR, engagement) rather than lagging ones (CPA, ROAS). If your link CTR is above 1% and people are adding to cart, the conversion metrics will likely improve as the algorithm learns. If your CTR is below 0.5% on day three with 2,000+ impressions, the creative is the problem and you can kill sooner.
For new Facebook ad campaigns, set a clear “testing budget” — the amount you are willing to spend to learn whether an ad works. A reasonable testing budget is 3-5x your target CPA per ad set. If your target CPA is $30, plan to spend $90-$150 per ad set during the testing phase. If the ad set uses up that budget without hitting your CPA target, you have your answer.
The Kill Decision Framework
Use this flowchart-style framework every time you evaluate an ad. Run through it top to bottom. The first “Yes” answer is your decision.
| Step | Question | If Yes | If No |
|---|---|---|---|
| 1 | Spent 2-3x target CPA with zero conversions? | Kill now | Go to step 2 |
| 2 | Still in learning phase (under 50 conversions, under 7 days)? | Wait — re-check in 2 days | Go to step 3 |
| 3 | Link CTR below 1% after 2,000+ impressions? | Kill — creative is the problem | Go to step 4 |
| 4 | CPA above 2x target for 5+ consecutive days? | Kill — audience or offer is wrong | Go to step 5 |
| 5 | Frequency above 3-4 with CTR declining? | Kill — ad fatigue, need new creative | Go to step 6 |
| 6 | ROAS below break-even for 7+ days? | Kill — not profitable | Go to step 7 |
| 7 | Two or more relevance diagnostics “Below Average”? | Kill — outcompeted on quality | Let it run |
If an ad passes all seven checks, it is performing well enough to keep running. Re-evaluate every 3-4 days using the same framework. Consistent application of this checklist eliminates emotional decision-making and keeps your ad account lean.
What to Do After Killing an Ad
Killing an ad is not the end of the process — it is the beginning of the next test. The data from a failed ad is valuable if you use it correctly. Here is what to do immediately after turning off a losing ad:
- Document what failed and why. Was it the creative (low CTR), the audience (high CPA despite good CTR), or the offer (low conversion rate on the landing page)? Each failure mode has a different fix. Do not just write “didn't work” — get specific.
- Reallocate budget to winners. The money you were spending on the dead ad should go to your best-performing ad sets. If you are already scaling your winners, increase their budgets by 20-25%. Do not let freed-up budget sit idle.
- Test one variable at a time. If the creative failed, test a new creative with the same audience. If the audience was the problem, test the same creative with a different audience. Changing both at once means you will never know which fix worked.
- Keep the winning elements. A failed ad often has one good component. Maybe the video hook was strong but the offer fell flat. Maybe the audience was right but the ad format was wrong. Extract what worked and build on it.
- Set a testing timeline. Launch your new variation and give it the same fair evaluation window — 7 days or 50 conversions, whichever comes first. Run it through the same kill decision framework. No shortcuts.
Common Mistakes: Killing Too Early vs. Not Killing Fast Enough
Most advertisers make one of two mistakes, and both of them are expensive:
Killing Too Early
This is the more common mistake, especially among newer advertisers. They launch an ad, see mediocre results on day two, and immediately shut it down. The problem is that day-two data is almost always noise. The algorithm has not had time to learn, the sample size is too small, and the CPA is unreliable.
If you kill every ad that looks bad on day two, you will never find a winner. Some of the best-performing ads in your account would have looked terrible if you judged them at the 48-hour mark. The learning phase exists for a reason — let it complete before drawing conclusions.
The cost of killing too early: you burn through creative concepts faster than necessary, never let the algorithm optimize, and end up thinking “Facebook ads don't work” when the reality is you never gave them a fair test.
Not Killing Fast Enough
This mistake is more expensive per occurrence. Advertisers who hold onto losing ads do so because of sunk cost bias (“I already spent $500, I should give it more time”) or because the ad used to work (“it was profitable last month, it will come back”). Both are emotional responses, not data-driven ones.
An ad that has been below break-even ROAS for two weeks is not coming back. An ad with a frequency of 6 and a CTR that has dropped by half is not going to recover. The data is clear. Every additional dollar you spend on it is a dollar you are choosing to lose.
The cost of not killing fast enough: you subsidize losing ads with profit from winners, your blended ROAS drops, and you miss the opportunity to test new ads that might actually work. True Margin exists to help you see these numbers clearly — your real margin after ad spend, not the vanity metrics that mask underperformance.
How to Find the Balance
| Mistake | What It Looks Like | How to Fix It |
|---|---|---|
| Killing too early | Turning off ads after 1-2 days or under 500 impressions | Wait for 50 conversions or 7 days minimum |
| Not killing fast enough | Running ads at 2x+ CPA target for weeks | Follow the kill framework — act on 5-day trends |
| Emotional decisions | “I feel like this one will turn around” | Only use the five metrics — no feelings |
| Sunk cost bias | “I already spent $800, can't stop now” | Past spend is gone — only future spend matters |
Know your break-even ROAS before you kill (or keep) a single ad.
Use True Margin's free ROAS calculator to find your exact break-even threshold, so every kill decision is backed by real math — not guesswork.
Open ROAS Calculator →The Complete Kill Thresholds at a Glance
Here is every kill threshold in one place. Print this out, bookmark it, or keep it open in a tab next to your Ads Manager. When you are staring at a campaign and wondering whether to cut it, run through this table:
| Metric | Kill Threshold | Minimum Data Required |
|---|---|---|
| Link CTR | Below 1% | 1,000+ impressions |
| CPA | Above 2x target | 5+ consecutive days post-learning |
| Frequency | Above 3-4 with declining CTR | Combined with performance decline |
| ROAS | Below break-even | 7+ days post-learning |
| Relevance Diagnostics | 2+ rankings “Below Average” | 500+ impressions |
| Zero conversions | Spent 2-3x target CPA | Any time (no minimum wait) |
The discipline of killing bad ads is what separates profitable ad accounts from money pits. Every dollar you stop wasting on a loser is a dollar you can invest in finding your next winner. The advertisers who grow their brands profitably are not the ones who never launch a bad ad — they are the ones who recognize it fast, cut it decisively, and redeploy the budget intelligently.
Frequently Asked Questions
How long should I wait before killing a Facebook ad?
Wait until the ad set exits Facebook's learning phase — typically 50 conversions or 7 days, whichever comes first. Killing an ad during the learning phase means you are judging incomplete data. The one exception: if the ad set has spent 2-3x your target CPA without a single conversion, you can kill it earlier because the signal is clear enough.
What CTR is too low for a Facebook ad?
A link CTR below 1% on a cold traffic campaign is a strong kill signal. It means your creative or targeting is not resonating. For retargeting campaigns, expect higher CTRs (2-5%). If CTR is below 1% after 1,000+ impressions, the ad is unlikely to improve without a full creative refresh.
Should I kill a Facebook ad with high frequency?
Kill the ad when frequency exceeds 3-4 and performance metrics (CTR, CPA, ROAS) are declining simultaneously. High frequency alone is not always bad — retargeting ads can run at higher frequencies profitably. But when high frequency pairs with rising CPA and falling CTR, that is textbook ad fatigue and the ad needs to be replaced with fresh creative.
What should I do after killing a Facebook ad?
Document what failed (creative, audience, or offer), reallocate the budget to your best-performing ad sets, and test a new variation that changes one variable at a time. Do not simply duplicate the dead ad with a higher budget — diagnose why it failed first. Common fixes include new hooks, different creative formats, or narrower audience targeting.
Is a low ROAS always a reason to kill a Facebook ad?
Not always. A low ROAS during the first 3-5 days of a new campaign can be normal while Facebook's algorithm is still learning. However, if ROAS stays below your break-even threshold for 7+ days with sufficient spend, that is a clear kill signal. Calculate your break-even ROAS using your gross margin before making any decisions — use our free ROAS calculator to find yours.