Why You Should Never Switch Off a Meta Ad

You open Ads Manager, sort by ROAS, and there it is: an ad with a 0.8x return sitting at the bottom of the list. Every instinct says switch it off and move the money somewhere better. It feels like good housekeeping.

It's one of the most common mistakes I see in Meta accounts, and it's quietly costing small brands sales. The way Meta delivers ads has changed, and the way we judge them needs to change with it. These days I almost never switch an ad off, and in this guide I'll explain why, and what I do instead.

The old way: treating your ads like a league table

For years, managing Meta Ads meant ranking ads against each other. Each ad had to earn its keep on its own numbers. Anything below your target ROAS got cut, and the winners got the budget. Simple, tidy and logical.

The problem is that it assumes each ad works alone: someone sees one ad, clicks it and buys. That isn't how people shop any more, and it isn't how Meta works any more either.

Why your account is an ecosystem

Think about how you buy something online yourself. You see a brand in your feed and scroll past. A few days later you see a video of the product in someone's home. You stop, look, maybe tap through to the site and leave. Then a week later you see a third ad, and that's the one you buy from.

Three ads did the work. Meta's reporting gives the credit to the last one.

The way I think about it: some ads are there to drive volume, and another ad confirms the sale. Your volume drivers fill the top of the funnel with the right people. Your confirming ads close the deal. Each one depends on the others.

So when you switch off the "bad" ad at the bottom of the list, you're often removing one of the ads that made your "good" ad look good in the first place. A week later the winner starts to slip, CPMs creep up, and it's hard to work out why.

Meta is already doing the sorting for you

Meta's delivery system has moved firmly towards automation: broad targeting, Advantage+ campaigns and machine learning that decides which ad to show to which person, and when. It works best when it has plenty of different creative to choose from.

That has a useful side effect. Meta already shifts budget away from ads that aren't working for a given audience. An ad it doesn't rate will naturally get a small share of spend. So that "underperforming" ad usually isn't eating much of your budget at all. It's sitting in the background, being shown to the people it suits.

When you switch it off, you aren't saving much money. You're just taking an option off the table for the algorithm.

What switching an ad off actually breaks

1. The customer journey

As above, the ads work together. Remove one stage and the others have to work harder, usually at a higher cost per purchase.

2. Learning and stability

Every significant edit to an ad set can push it back into the learning phase. Turning ads on and off every few days keeps the account in a constant state of reset, and performance never settles.

3. Social proof

An ad that's been live for weeks has built up likes, comments and shares. That engagement makes it more persuasive to the next person who sees it. Switch it off, and launch a replacement, and you start again from zero.

4. Your retargeting pools

Cheap volume ads are often what keeps your video viewer and site visitor audiences topped up. Cut them, and a few weeks later your retargeting has fewer warm people to show to.

So how should you judge an ad instead?

ROAS still matters, but it isn't the way to judge an individual ad, and it shouldn't be the only way you judge the account. Here's what I look at instead:

What to check	Why it matters
Account-level results	Total Meta spend against total Shopify revenue for the period. If that number is healthy, the ecosystem is working, whatever individual ads say.
Share of spend	If Meta is barely spending on an ad, it isn't costing you much. If it's spending heavily, Meta has found a job for it.
Hook and hold	Does the video stop the scroll and keep people watching? That tells you whether it's doing its job as a volume driver.
CPM and reach	Cheaper reach into new people is exactly what you want from top of funnel ads.
Comments and saves	Signs of genuine interest, and social proof that helps every future impression.

What to do instead of switching ads off

Add, don't subtract

If performance is slipping, the answer is nearly always new creative, not fewer ads. Give Meta fresh angles to test: a new hook, a different format, a customer video, a product close up. Let the system find where each one fits.

Refresh on a rhythm

I add new creative every few weeks rather than waiting for things to drop. It keeps the account fresh without the stop-start of cutting ads.

Move winners up, don't kill the rest

When an ad clearly proves itself, duplicate it into a scaling campaign or ad set with more budget behind it. The original stays where it is, still doing its job.

Make changes slowly

Change one thing at a time and give it time to settle. Budget changes of around 20% at a time are far less disruptive than big swings.

Judge the account weekly, not the ads daily

Day to day numbers on individual ads are noisy. Step back and look at the whole account over a week or more before you decide anything.

The exceptions: when you should switch an ad off

"Never" is a rule of thumb, not a law. There are a handful of times when turning an ad off is the right call:

• The product is out of stock or discontinued, and the ad points straight to it.
• The offer has ended, so the ad is now promising something you no longer do.
• The link is broken or the landing page has changed.
• The comments have turned negative and the ad is doing your brand more harm than good.
• It's out of season, like a Christmas ad still running in February.
• It's genuinely eating budget with no sign of reach, engagement or sales over a sensible period, and the account as a whole is struggling.

Notice what's not on the list: "it has a low ROAS".

A quick checklist before you touch anything

• Have I looked at total spend against total Shopify revenue, not just ad-level ROAS?
• Is this ad actually taking a meaningful share of the budget?
• Is it reaching new people cheaply, or getting strong watch time and engagement?
• Have I given it at least a week, ideally longer, before judging it?
• Would adding new creative solve this better than removing an ad?
• Does it fall into one of the genuine exceptions above?

If you can't tick the last one, leave it running.