Turn off a Facebook ad when it has spent twice what your best ad pays for a lead, and it is still paying twice that or worse. That is the whole rule. We call it the 2x rule, and our team runs it on every live ad we manage.
Most owners decide when to turn off a Facebook ad by feel. An ad looks busy, so it stays. An ad brought in one great customer, so it stays. An ad looks quiet on a Tuesday, so it goes. Each call feels fair in the moment. Together they leave your budget spread across ads that cost far more than they should.
The fix is to stop judging ads one at a time. Every ad gets judged against your best ad, in one table, with the same math. Below is each part of the rule, what it costs you when that part breaks, and what good looks like.
The table that puts every ad side by side
For every live ad, we write one row. It shows what the ad has spent, how many leads it brought in, its cost per lead and its multiple of the best ad. Cost per lead is the price you pay for each form fill. The multiple shows how many times more this ad pays for a lead than your best ad does.
We call it an opportunity cost table. Every dollar on a weak ad is a dollar your best ad did not get. So the question is never "is this ad bad?" The question is "what would this money have bought on the best ad?"
What it costs you when it breaks: without the table, each ad gets judged alone. A lead price that looks fine on its own can be three times what your best ad pays. Owners also keep the ad they like most, even when it costs the most.
What good looks like: every live ad is on the table at every read. No ad gets left off because it is new or because someone likes it. The worst ads sit at the bottom with their multiple in plain view.
Picking the best ad
The best ad is the cheapest ad with 3 or more leads in the same funnel. A funnel here means one offer, going to one form and one booking page. Two rules hide in that sentence.
The first is 3 or more leads. One cheap lead can be luck.
The second is the same funnel. Different offers bring in leads at different prices. A free consult and a paid session will rarely cost the same per lead. Comparing them tells you about the offers, not about the ads.
What it costs you when it breaks: pick the wrong best ad and every multiple on the table is wrong. Set the bar with a lucky ad and you cut good ads that are just normal. Compare across offers and you cut an ad for selling something harder.
What good looks like: one best ad per funnel, with enough leads behind it to trust. When another ad beats its cost per lead with 3 or more leads of its own, that ad becomes the new bar.
The 2x cut line
An ad gets cut when two things are true. It has spent at least twice the best ad's cost per lead. And it is paying 2x the best ad's cost per lead or worse.
Here is a made-up example to show the math. Say your best ad brings in leads at $20 each. Any ad that has spent $40 has had a fair shot. If it has no lead yet, it comes off. Another ad has spent $120 and brought in two leads, so it paid $60 per lead. That is 3x the best ad, so it comes off too. A third ad spent the same $120 and brought in four leads. It paid $30 per lead, or 1.5x. It stays.
The spend half protects new ads, since nothing gets cut before it could have won one lead at your best price. The multiple half protects your budget.
What it costs you when it breaks: without a line, cuts come late or not at all. Each weak ad keeps its slice of the budget, while the best ad waits for money it never gets.
What good looks like: the line is written down before the ads go live. Nobody argues it ad by ad. When an ad crosses it, the ad comes off that day.
What never saves an ad
Two things tempt owners to keep a weak ad on. One is a closed deal. The other is a good click-through rate, the share of people who click after they see the ad. Neither one saves an ad that has crossed the line.
We learned this with our own money. We run ads to reach portrait studio owners. From 10/9 to 10/12, our best ad there ran about $26 per lead. Another ad ran $199 per lead, which is 7.6x. We kept the expensive one on twice because it had won a deal. In those 3 days it burned $253 and brought in no lead.
The deal was real, but it was in the past. It told us the ad found one buyer once. It did not tell us the next dollar would find another.
A good click-through rate is the same trap. Clicks show that people are curious. They do not show that people act. And if all your ads get strong clicks with zero leads, look at the page or the form first, not the ads.
What it costs you when it breaks: every day a favorite ad stays on, it takes money from a cheaper ad.
What good looks like: a paused ad stays paused. We never turn a worse ad back on. If you miss its message, write a new version of the best ad that carries it.
The dry spell rule
Some ads start strong and then go quiet. Their total cost per lead still looks fine because of the early leads. So the table tracks one more number: spend since the last lead.
Our rule is simple. An ad that has spent $150, or 2x the cost per lead, since its last lead gets cut, even if its total looks fine.
What it costs you when it breaks: an ad can coast for days on old results. The total hides the fact that its recent money bought nothing.
What good looks like: spend since the last lead sits next to the totals on every row. A dry spell shows up the day it starts, not a week later.
Where the money goes after a cut
Cutting is half the move. The money a cut frees up goes to one of two places: new versions of the best ad, or ads that have never been tested. It never goes back to an ad that already lost.
A new version keeps what works about the winner and changes one piece, like the hook or the photo. We never edit a live ad to do this. A copy change makes a new ad and starts its learning over.
Where those new ads sit depends on your budget. An ad set is a group of ads that shares one budget and one audience. Under $100 a day, new ads go into the winner's ad set and get judged over a full week. At $100 a day or more, we clone the winner's ad set and move about 10% of the budget into it, so the total stays flat. Below about $100 a day, a separate test set starves both the test and the winner.
What it costs you when it breaks: if cuts free up money and nothing new is ready, the whole budget piles onto one winner. Every winner wears out in time. So we keep the winner's next versions ready before the winner is about 7 days old.
What good looks like: every cut has a replacement waiting.
The caveat: low spend and early reads
The 2x rule needs enough money behind each ad to mean something. At low spend, single-ad verdicts are mostly noise. When each ad has spent $30 to $75 against about a $20 cost per lead, each ad holds 1 to 4 leads. That is not enough to rank them. In our simulation, a rule of "$30 and no lead means kill" would cut about 22% of ads that were doing fine. So at low spend we judge groups of ads, not single ads.
Timing matters too. A zero-lead first day is normal on a new pixel, the tracking code that tells Meta who became a lead. We act at 72 hours and about $300 spent, not before. Early on, one ad often takes most of the spend. That is the auction picking a winner. Before 72 hours, we pause a single ad only for a defect: a wrong link, a broken creative or a policy problem.
We also read new campaigns by the hour, not just by the day, and at the ad set level over 7 to 10 days. An ad that looks dead by the day can be fine by the hour.
How the parts fit together
Here is the whole rule on one account. A boudoir studio in the Midwest runs a giveaway offer. Before our team took over, its leads cost $39.88 each. Our first read after the relaunch showed $1,279 for 92 leads, or $13.91 per lead. That same day, 14 ads came off under the 2x rule.
Three days later we read the table again. Two more ads had crossed the line, at 2.1x and 2.6x the best ad's cost per lead. They came off. We also raised the budget on the best audience from $50 to $150 a day.
That is the loop. Build the table, find the best ad and cut what crosses the line. Watch for dry spells, move the money, then read again.
One last limit. Cost per lead is the fastest number to read, so it runs the table. It is not the final judge. A cheap lead that never books is still an expensive lead, which is why we judge ads in the end on the cost of a booked job. And as we covered in why your Meta cost per lead can mislead you, the lead number is only as honest as the tracking behind it. The table's job is to stop the obvious waste fast, so the money goes where closed revenue comes from.
FAQ
How long should a Facebook ad run before you turn it off?
Give a new campaign 72 hours and about $300 in spend before you act. A zero-lead first day is normal. After that, an ad comes off once it has spent twice your best ad's cost per lead and is paying 2x or worse. The only early cuts are for defects, like a wrong link or a broken creative.
Should you keep an ad that brought in a sale?
Not if it has crossed the 2x line. A past sale shows the ad found one buyer once. Our own ad at $199 per lead had won a deal, and keeping it on cost us $253 in 3 days with no lead. Put that money behind your best ad instead.
Can you turn a paused ad back on later?
We never turn a worse ad back on. If you liked something about it, build a new version of your best ad that carries that idea. That new version gets its own fair test against the line.
What if none of your ads have spent much yet?
Then judge them as a group. At low spend, each ad holds only a few leads, so single-ad verdicts are mostly noise. Wait until the campaign has real spend behind it, then build the table.
If you want a second set of eyes on which of your ads should stay on, book a discovery call with our team.