Meta Ads Placements: Where Your Budget Leaks

Manual placements settings in a Meta Ads ad set

Open any Meta campaign that was set up with the default settings and add one breakdown to the table. You will find spend in places you never chose: apps you have never heard of, video slots inside other people's content, the search results page, the right-hand column on desktop. Some of those placements bring purchases. In the sales accounts I run, most of them bring impressions and clicks and little else.

The fix takes five minutes in the ad set settings, and a check a week later shows whether it held.

The breakdown by placement

In Ads Manager, open the campaign, set the date range to the last 30 days, and click Breakdown above the table. Choose By delivery, then Placement. Each ad set splits into one row per placement: Facebook Feed, Instagram Feed, Instagram Reels, Instagram Stories, Audience Network, Messenger, Marketplace, In-stream video, Search results, Right column, and so on.

Read three columns for each row: amount spent, purchases (or leads), and cost per purchase. In most sales accounts I have looked at, the feeds, Reels, Stories and Explore carry the purchases, while Audience Network and the smaller placements show a low cost per click and almost no purchases. Write down the share of spend that went to placements with no purchases, so you can compare it after the change.

Manual placements, and the checkbox under the list

Open the ad set, scroll to Placements, and switch from Advantage+ placements (Meta's automatic choice) to Manual placements. Expand the full list and untick everything except the placements that carried purchases in your breakdown. For a store selling to people who discover products while scrolling, my list is Facebook Feed, Instagram Feed, Instagram Reels, Instagram Stories and Instagram Explore.

Then look for a checkbox under the placement list that lets Meta spend part of the budget on the placements you just excluded. According to reporting on the feature, it is labelled "Allow limited spend to excluded placements", appears for sales and leads campaigns, is switched on by default, and allows up to 5% of budget per excluded placement. Source: PPC Land, October 2025. Meta changes labels from time to time, so read the wording in your own account. That checkbox is easy to miss: you can untick nine placements and still fund them, a few percent each, because the box under the list stayed ticked. Untick it, publish the change and note the date.

Meta's argument for automatic

Meta's interface marks the automatic option as recommended, and its reasoning is consistent: more placements mean more auction opportunities and, on average, a lower cost per result. That average includes cheap results that are not purchases, so a lower cost per result can come with a higher cost per purchase. When the goal is a sale of a physical product, the placements that produce those sales are a short list, and the rest dilute the budget. I would rather pay a little more per impression in the feed than pay less for taps in an app I cannot name.

There are cases where I leave the automatic setting on: a campaign whose goal is reach or video views, where the cheapest impression is the point, or an account with a very large budget that has already saturated the feeds. For a sales campaign with a normal budget, manual has won in the accounts I have touched.

Seven days later

Run the same breakdown by placement for the seven days after the change. The rows you excluded should show zero spend; if one still shows spend, the "allow limited spend" box is ticked somewhere, or a second ad set kept the old setting. Compare cost per purchase before and after. Expect the cost per click to rise a little, because the cheapest clicks are gone, and the cost per purchase to fall or hold. Check reach as well: if it dropped sharply and purchases fell with it, one of the placements you removed was doing real work. Add that one back and leave the rest excluded.

What the leak usually looks like

In the accounts I have audited, the leak is many small placements each taking a little, together something like 20% of the budget, going to places that bring no purchases. Manual placements close all of them at once, and the ads themselves do not change, only the places they run.

My tip: I don't hand placement choice to Meta. For sales, the feeds, Reels, Stories and Explore do the work, and everything else in my accounts only spends the budget. Five minutes in the ad set settings save a noticeable part of the spend every month.

If you are not sure what your breakdown is telling you, or you want the whole account checked for leaks like this one, book an audit.

Read next

Lookalike audience vs broad targeting on Meta ads
Meta Ads

Lookalike Audience vs Broad Targeting On Meta: What Worked After $3M In Ads

Most advice says go broad on Meta. After almost $3M in Meta ads, I still start with interests, add lookalikes second and test broad last. Here is how I split the budget.

Nov 2, 2026 Read →
Meta Ads reports built from the Marketing API
Meta Ads

Your Own Meta Ads Reports: How I Connect To The API Without Paid Dashboards

Ads Manager has every number, in Meta's layout. A small app of your own puts the six numbers that matter in one row every week. Setup, tokens, and what one report found.

Sep 25, 2026 Read →

Want an Audit?

Contact Me
Book a Call