Three arguments. One answer.

Why the market misses this audience, how we reach it instead, and why the same design decision that makes us accurate is what makes us private.

Since 2015built for U18, not retrofitted
250K+households tracked in 7 markets
Zeropersonal data collected

01 · The U18 gap

The reach is real. The buying is broken.

Seven of every ten dollars aimed at the under-18 audience never reach it, what happens when an audience is bought through systems designed for someone else.

Path A · general DSPs

Built for adults, filtered for kids

  • Age filters bolted on after the fact, with no behavioral signal under 18, the audience is a late idea in the product.
  • Age inferred from proxies that cannot distinguish a six-year-old from a sixteen-year-old.
  • Buying U18 on YouTube by interest sends most of the spend into adult content, more of it when teen keywords are used.

7 of every 10 dollars never reach the intended audience.

Path B · fragmented kids vendors

Built for kids, one slice at a time

  • One platform, one channel, one format, one age band, contextual activation only, with no omnichannel planning.
  • Duplicated cost across vendors, and data that stays fragmented.
  • Five contracts, five reports, and no way to reconcile them.

No single answer to who you actually reached.

02 · Audience-behavior-first

Everyone classifies the content.
We classify the audience.

Content-first

Choose the media, then hope

  • Classify the page, the keyword, the video.
  • The audience is whoever happened to be there.
  • Safety is a blocklist added after the media was chosen.
  • Reach is inventory reach, it can't be deduplicated or verified against delivery.
Kidscorp

Define the audience, then find it

  • Every channel, app and content unit classified by what U18s declare they consume.
  • Audience defined first: age band, gender, market, passion points, household co-viewing.
  • The model predicts where that audience is, and the bid lands there.
  • Reach figures are audience figures, deduplicable across YouTube, apps and CTV, and verifiable against delivery by age once the campaign has run.

We don't identify the child. We profile the placement.

The consequence

This is why the privacy question has a different answer here.

Accuracy and privacy are usually a trade. Here they are the same decision, taken once, at the level of what the model looks at.

Nothing to identify

Because the unit of analysis is the placement, not the person, there is no profile of a child to build, store or lose.

Nothing to lose

Because the model never used identifiers, none of it degrades as the industry's identifiers disappear.

Nothing to explain away

Contextual activation built on declared U18 consumption is both the accurate option and the compliant one.

03 · Privacy architecture

Private by architecture, not by policy.

Most platforms describe privacy as promises about the data they collect. We never collect it: the model was built to work without personal data from a minor, so compliance is a property of the system rather than a commitment layered on top.

01

No PII from minors

No personal data is collected from anyone under 18, anywhere in the platform.

02

No user-level tracking

No device graph, no cross-site identifier, no individual profile.

03

No retargeting

A minor is never followed from one impression to the next.

04

No client data on targeting

No first-party, CRM or matched audiences layered onto U18 targeting.

05

Households, not users

Panel signal is relayed from consenting households with parental authorization, never inferred from device behavior.

Regulatory map

What the architecture already satisfies.

These frameworks govern the collection and processing of children's data. A platform that collects none of it meets them structurally.

FrameworkScopeHow the architecture answers it
COPPA United States, collection of personal information from children under 13 No personal information is collected from any minor, nothing to obtain consent for, store or disclose.
U.S. state laws State children's and teen privacy provisions No profiling of minors, no targeted advertising on personal data, no sale of minors' data, none of which the platform performs.
GDPR-K European Union, processing of children's personal data No processing of children's personal data; panel participation is separately consented at household level.
LGPD Brazil, dedicated chapter on children and adolescents Contextual activation on declared consumption, with no individual profiling of minors.

Brand safety

Automated filters plus human moderation against the U18 content model. Blocklists apply at model level, before the bid, not as a post-bid filter on media that already ran.

Creative review

Every asset is reviewed for compliance before it is eligible to serve, and produced against age-band UX and accessibility requirements.