The YouTube Kids Channel Cleaner Grows Up: PMax, Countries and Languages

In 2017 I noticed something absurd in the YouTube campaigns I was running: in some accounts, up to 90% of impressions targeted at adults were actually being shown to children — kids watching cartoons on their parents’ devices, happily tapping whatever appeared on screen. The engagement metrics looked wonderful. The sales effect was zero: a four-year-old does not buy your product, no matter how many pre-rolls they sit through.

So I wrote a script. The useful discovery back then was that children’s channels are grouped not by their names but by what their videos say about themselves — the same kids-content keywords repeat across channels in every language. The script checked every new placement against a multi-language keyword library and pushed matching channels into an excluded placement list. I documented the whole approach in an article for Semrush, and I knew it was working when the view-through rate went down: adults press Skip far more often than children do. A falling VTR meant the budget was finally reaching people who could actually say no — and, more importantly, people who could say yes.

Nine years later that script — now the YouTube Kids Channel Cleaner in the chiliad marketplace — just got its biggest update since the original. And this time the changes are not driven by a hunch from a handful of accounts. They are driven by research that measured the problem across billions of dollars of ad spend.

2026: The Problem, Measured at Scale

In June 2026, Wijnand Meijer at TrueClicks published a study that did what no single advertiser could: it analyzed placement data from more than 40,000 Google Ads accounts representing over $5 billion in yearly spend, reviewed nearly 4 million YouTube channels, and compiled a free, quarterly-refreshed list of 30,818 channels targeting children.

The study confirms what the 2017 experiment could only suggest — and puts numbers on it:

• Kids traffic is worse than useless. Children cannot buy what you sell, their engagement is inflated by accidental taps and autoplay, and for age-gated industries — alcohol, gambling, finance — serving them ads is a compliance risk on top of the wasted budget.
• You cannot trust the label. TrueClicks classified channels by their actual content, not by YouTube’s “made for kids” flag, because the flag alone misses too much. A serious cleaner needs multiple independent signals.
• 46% of kids channels are not in English — the list spans 100 languages. A filter that only speaks English is blind to nearly half the problem.

Or, as Meijer puts it: automation may have taken the controls out of your hands, “but not the responsibility for where your money lands.”

What the Research Demanded from the Script

This update maps each finding onto a capability. Three big ones.

1. Language and Country Filters — Because Half the Problem Does Not Speak English

If 46% of kids channels are non-English, a keyword library alone will always lag behind — you cannot maintain blocklist keywords in 100 languages. So the script now reads each placement’s language and each channel’s country and applies blocklists or allowlists directly:

• BLOCKED_LANGUAGES / ALLOWED_LANGUAGES — exclude specific languages (e.g. ['ru', 'ar']), or the stricter inverse: exclude every language not in your list (e.g. ['en', 'nl']).
• BLOCKED_COUNTRIES / ALLOWED_COUNTRIES — the same pair for the channel’s country, using standard two-letter codes.

One deliberate rule: placements with an unknown language or country are never excluded by these filters. Missing metadata is not evidence of anything, and a brand-safety script should not punish it — those placements land in the review sheet instead, where a human decides.

2. Performance Max Coverage — Because PMax Ignores Your Placement List

The original script had one assumption baked in: if a channel is on your excluded placement list, Google stops serving ads on it. For video and Display campaigns, true. For Performance Max — false. PMax ignores placement exclusion lists entirely, and its placements live in a separate report the script never looked at. Two blind spots stacked: the script could not see PMax placements, and even if it could, the list would not have gated them.

The only exclusion level PMax honors is the account level. So that is where the script now operates:

Two new config variables control this, both on by default:

• INCLUDE_PMAX — the script now scans the PMax placement report alongside regular video placements. Channels flagged there are excluded account-wide immediately, because a placement-list entry would be ignored. One thing to know: the PMax report only shows impressions, not cost, so PMax-only rows show 0 in the Cost column of your review sheet, and PMax placement data can lag a day or two behind.
• SYNC_LIST_TO_ACCOUNT — this one heals history. Every run, the script mirrors your entire exclusion placement list into account-level exclusions: add-only, idempotent, whitelist respected. If you have been running the cleaner for years, every channel it ever excluded starts gating PMax on the very next run. Nothing to migrate, nothing to re-flag.

3. Spend the 65,000 Slots on Channels That Actually Serve

Google caps account-level exclusions at 65,000 per account, and TrueClicks makes the arithmetic explicit: paste their full 30,818-channel list into a single account and you have burned nearly half your exclusion capacity on channels that may never serve your ads at all. Their own recommendation for individual accounts is to use a script that compares the list against your actual placements and excludes only the channels that really appeared.

That has been this script’s design since 2017. It never excludes speculatively — it watches what your campaigns actually serve on, checks each new placement against every signal (kids flag, keywords, language, country, blocklist), and excludes only confirmed offenders. Your 65,000 slots go to channels that were demonstrably costing you money, and the script logs a hint if you ever approach the cap.

Using the 30,818-Channel List with the Cleaner

The static list and the dynamic script are not competitors — they compose. The cleaner maintains a blocklist sheet in its spreadsheet: any channel ID or channel URL you put there is excluded the moment it shows up in your placements, no questions asked. Paste TrueClicks’ list into that sheet and you get the best of both approaches:

• Known kids channels are excluded on first contact — before the keyword, language or kids-flag checks even need to run.
• No capacity is wasted: a channel from the list only consumes an exclusion slot if it actually appears in your traffic.
• With SYNC_LIST_TO_ACCOUNT on, those exclusions land at account level too — so the list gates Performance Max, which a pasted placement list on its own never would.
• When TrueClicks refreshes the list next quarter, paste the update into the same sheet. Done.

And everything still lands in the audit trail — the part of the 2017 design I have never changed. Every processed placement is logged to a review sheet with its language, country, Made-for-Kids flag, YouTube category, impressions, cost, and now a PMax column. Sort by impressions, filter by language, and you will spot the next batch of exclusion candidates before they cost you anything. A whitelist sheet protects channels from ever being excluded (the account-level mirror respects it too), and Preview mode shows exactly what would happen without writing a single file or applying a single exclusion.

Nine Years On

The 2017 script solved the problem of one era: cartoons soaking up video budgets on parents’ devices. The funny thing about brand safety on YouTube is that the problem never goes away — it changes shape. Placement lists were the answer until Google shipped a campaign type that ignores them. English keywords were the answer until research showed that nearly half the problem does not speak English.

What the TrueClicks study really proves is that this is not one advertiser’s quirk — it is a structural, $5-billion-wide leak, and it needs both a map (the 30,818-channel list) and a mechanism that acts on your account’s real traffic every day. The updated script is that mechanism: same promise as 2017, three more tools to keep it.

The YouTube Kids Channel Cleaner is available in the chiliad marketplace, alongside 120+ free Google Ads scripts. If you already run it through chiliad, you will see an “Update available” banner on your installed copy — one click upgrades it while preserving every variable you have customized. And if you set it up in 2017 from the Semrush article: welcome back, your exclusion list is about to start working on PMax.