Kantar/NMO: how is an app visit counted as audience measurement?

Who does what?

NMO, Nationaal Media Onderzoek, coordinates Dutch media audience research. Kantar carries out its online and viewing research. NMO describes a combination of recruited panels and census measurement: tags in participating websites and apps count use. A tag is measurement code; being counted by a tag is not the same as agreeing to participate in a research panel. NMO participant information.

NMO's FAQ says census data include pageview counts and characteristics such as duration, content category and device. Those counts contribute to reach estimates. This describes the research model, rather than the precise contents of every mobile request. NMO Online FAQ.

Why can a counting request contain an ID?

Hypothetical example: a service wants to distinguish ten opens by one installation from one open by ten installations. A repeated identifier helps it recognise recurring measurement records. An identifier does not have to contain a person's name to connect activity. Conversely, finding such a value does not prove that a research organisation knows the name of the person holding the phone.

A ping is a small network request reporting a measurement. It can have parameters: named values attached to that request. What those parameters mean must come from the implementation, rather than from the harmless-sounding word “counting”.

What did we find in Buienradar?

In Buienradar 7.4.8, a plugin attached to the RTL Snowplow tracker adds a Kantar/NMO ping to nmortlendpoint.2cnt.net. The parameter cs_fpid receives the Snowplow user ID. This is concrete evidence of an identifier connection in the app's code. It is not a packet capture showing a particular person's request, and does not establish what downstream matching, retention or aggregation occurred. Pinned Buienradar audit.

What changes after patching?

Buienradar's Disable tracking prevents creation of the Snowplow tracker to which this plugin belongs. The identified NMO ping stops with that tracker. This effect depends on the inspected integration, rather than on blocking all domains used by a research company.

Do not apply that conclusion automatically to another app or a later release: the same audience scheme can be integrated through independent code. Nor should NMO's panel description be used to dismiss the separately documented census tag. Our evidence describes the app-side path; the research provider's documentation explains its stated purpose.

Evidence reviewed 28 October 2026. Read identifiers, consent and evidence levels.