Meta Ads · 7 min read
How to Stop Meta Wasting Your Budget on Audience Network
There is one leak that quietly drains more small Meta budgets than almost any other, and many operators never see it because their dashboard looks healthy. Cheap CPM, plenty of video views, low cost per ThruPlay. Green across the board. Meanwhile much of that spend may never reach a real person who might buy from you.
A common culprit is Advantage+ Placements, Meta's default setting that lets the system decide where your ads run. Left on, it optimizes for the cheapest impression it can find, and the cheapest impressions are often on the Audience Network, particularly rewarded video inside mobile games where someone watches an ad to unlock a life or a coin. This guide shows you how to find that leak, stop it, and keep it from coming back after future edits.
Why Advantage+ Placements can leak your budget
When you create a new ad set, Meta defaults to Advantage+ Placements, which means it can deliver your ad across Facebook, Instagram, Messenger, and the Audience Network. The system optimizes for your stated goal at the lowest cost, and it often learns that the lowest-cost impressions are off-platform. So it can shift budget toward whatever is cheapest, and that is frequently the Audience Network.
The slice to watch most is rewarded video. These are the ads inside free mobile games where a player watches in exchange for an in-game reward. Many of those viewers are not choosing to engage with your offer, they are tapping through to get back to their game. The ThruPlay can look real and the cost per view can look fantastic, but there may be little intent behind it. The result can be a campaign that posts attractive surface metrics and very few landing-page views or conversions.
This is not a rare edge case. On one real account, roughly 97 percent of spend had been flowing to Audience Network rewarded video, and about 87 percent was concentrated in a single country that was never the target, before anyone caught it. Everything upstream looked fine. Those exact figures come from one account and are not a typical or guaranteed outcome, but the pattern they illustrate is common: the money quietly going somewhere useless while the dashboard stays green.
How to detect the leak with a placement breakdown
You usually cannot see this in the default summary view, because it aggregates across placements. You have to break the data apart. Pull a performance report broken down by publisher platform and platform position, and pull a second report broken down by country.
The dimensions to use are publisher_platform (the network: facebook, instagram, audience_network, messenger), platform_position (the specific surface, for example feed, instream_video, marketplace, search, right_hand_column, and the rewarded-video surface inside Audience Network), and country. In the Meta Marketing API these are values for the breakdowns parameter on the insights endpoint. In Ads Manager you reach the same view through the Breakdown menu, by placement and by country.
Read it like this. Any spend on publisher_platform equal to audience_network should usually be near zero unless you deliberately want it there. Watch specifically for the rewarded-video and in-stream video positions. On the country report, watch for one country taking far more or far less than its fair share of spend, since a wildly imbalanced split can be a sign that Meta is chasing cheap impressions instead of your real audience. If most of your budget is on Audience Network or piled into the wrong geo, you have likely found the leak.
How to lock placements to Facebook and Instagram
The fix is to switch from Advantage+ Placements to manual placements and explicitly exclude the surfaces that drain budget. Keep delivery on Facebook and Instagram feed, reels, and stories, where your audience is actually choosing to scroll. Exclude the Audience Network, and in most cases also exclude Messenger, in-stream video, search, right-hand column, and marketplace, since those often carry less intent for a direct-response funnel.
In Ads Manager this is the Manual placements option inside the ad set, where you uncheck everything except the Facebook and Instagram feeds, reels, and stories you want. Through the API you set the publisher_platforms field along with facebook_positions and instagram_positions in the ad set targeting. The point is the same either way: you, not the system, decide where the money goes.
This single change is often the difference between a campaign that looks busy and a campaign that works. It will usually raise your CPM, because real feed placements tend to cost more than rewarded video, and that is expected. You are paying more per impression to reach people who can actually convert.
The post-edit re-enable trap
Here is the part that catches even experienced operators. Locking placements once does not guarantee they stay locked. If you later patch an ad set's targeting, for example to change the audience or geo, and your update omits the publisher_platforms field, Meta can quietly switch placement control back toward Advantage+ Placements. The leak can come back without any obvious warning, and you may not notice until you break down the data again.
So treat every targeting edit as a moment to re-verify. After any change to an ad set, pull the placement breakdown again and confirm the Audience Network is still near zero. If you edit through the API, include the publisher_platforms field in the patch even when you are only changing something else, so you never hand placement control back to the system by accident.
The broader lesson is that a green-looking dashboard is not evidence the funnel works. Verify the wiring, not just the summary numbers.
The 7-minute daily check that catches leaks early
A short, repeatable routine keeps this and similar problems from compounding. Once you know the account, the whole thing can take roughly seven minutes. Step 0, funnel integrity: confirm the live audience and creative wiring actually matches what you think is running, because surface metrics looking green is not proof the funnel is connected. Step 1, account health: scan for disapprovals or delivery issues. Step 2, yesterday's performance against your targets. Step 2b, the placement and geo audit covered above, which is the most important step. Step 3, audience growth. Step 4, downstream signals like site traffic, signups, and whether your conversion events are actually firing. Step 5, a short status report so you have a running record.
A few rough benchmarks help you read the numbers, but treat them as ranges that vary a lot by niche, geo, objective, and season, never as universal truths. For many feed and reels campaigns, CPM often lands somewhere in the low tens of dollars, though plenty of accounts sit well above or below that. Cost per ThruPlay of a few cents can be fine for genuine viewers, while a suspiciously low sub-cent figure can be a sign of a rewarded-video leak. A landing-page-view rate above roughly 50 percent is often considered healthy, while below about 25 percent can suggest a broken landing page or the wrong audience. Frequency climbing past roughly 4 per week on a feed or reels placement can risk ad fatigue, which is a cue to consider rotating creative. Think of the funnel as a pipeline, from impressions to reach, to 3-second and then 25, 50, 75, and 100 percent video views, to landing-page views, to add-to-cart or lead, to purchase or subscribe. A weak hook tends to show up as people dropping between the 3-second and 25 percent views. A broken landing tends to show up as clicks and views that never become landing-page views. A leak at any stage starves the next.
One practical note if you build video-view retargeting audiences from this funnel: Meta exposes specific watch thresholds (for example 25, 50, 75, and 95 or 100 percent, plus a short 3-second or ThruPlay-style watch event), and the available options can differ between the Ads Manager UI and the Marketing API. Pick the threshold from the options Meta actually offers in the surface you are using rather than assuming a custom percentage will be accepted. Doing this whole check by hand every day is tedious, which is exactly why it tends to get skipped right when it matters most. An agent like AdsBud can run it for you: it reads the account around the clock, scores each campaign against today's number, your 7-day average, and your since-launch baseline, flags leaks like the Audience Network drain, and proposes a fix for one-click approval. It is read-only by default with one-click rollback, and you talk to it in your own words, so the daily check happens whether or not you have seven minutes to spare.
Frequently asked
Should I ever leave the Audience Network on?
Sometimes, but it is often a poor fit for a small direct-response budget. If you have tested it deliberately and it produces real landing-page views and conversions at an acceptable cost, keep it. The problem is not the Audience Network existing, it is Meta defaulting your budget into its cheapest, lowest-intent corner (often rewarded video) without you choosing it. Decide on purpose, then verify with a placement breakdown.
Won't manual placements hurt my reach and raise costs?
Your CPM usually does go up, because real Facebook and Instagram feed placements tend to cost more than rewarded video. That can be a worthwhile trade. You are exchanging a cheap, low-intent impression for a more expensive one that can actually convert. Reach quality within your true target audience often improves even if the raw impression count drops. Judge it by landing-page views and conversions, not by CPM alone.
How do I know if low cost per ThruPlay is good or a warning sign?
Context decides it. A few cents per ThruPlay from genuine feed and reels viewers can be healthy. A suspiciously low sub-cent figure, especially paired with very low landing-page-view rates, is a common fingerprint of a rewarded-video leak where people watch without engaging. Always cross-check cheap video metrics against the placement breakdown and your downstream conversion events before trusting them, and remember these are rough ranges that vary by niche and objective.
I locked my placements weeks ago, why is spend on Audience Network again?
Often this is a post-edit re-enable. If you patched the ad set's targeting and the update omitted the publisher_platforms field, Meta can switch placement control back toward Advantage+ Placements. Re-apply manual placements, and from now on re-pull the placement breakdown after every targeting edit. When editing through the API, include the publisher_platforms field even when you are changing something unrelated.
See what AdsBud catches on your account
AdsBud runs this check for you around the clock, flags the leaks, and proposes each fix for your one-click approval. Read-only by default, cancel any time.