Will We Still Conduct Surveys and Audience Measurement in 2030? Collecting Data When Everything Stands in the Way

Reading time
2 mn 51 s

By Yannick Carriou, President and CEO of Médiamétrie

 

14th Francophone Symposium on Surveys, May 19–22, 2026, Vannes

Image
Yannick CARRIOU

Abstract

Telephone contactability has become one of the most severe constraints on the production of representative data. Over the past decade, response rates in telephone surveys have collapsed across industrialised countries. In the United States, the Pew Research Center documents a fall from 36% in 1997 to less than 6% in 2018, ultimately leading the institution to abandon outbound telephone polling entirely in 2022. In France, Médiamétrie's survey operations record historically low pick-up rates, while immediate refusals and calls terminated before any argumentation can be made continue to rise. This deterioration is not cyclical: it reflects the convergence of three structural dynamics.

The first is behavioural. A growing fraction of the population deliberately refuses to answer calls from unknown numbers, driven by distrust of commercial canvassing and telephone fraud. The ARCEP Annual User Satisfaction Survey (2025 edition) finds that over 70% of mobile users encounter fraud or unwanted calls on a weekly basis, leading a majority of them not to answer unrecognised numbers. This heightened vigilance translates into growing selectivity that penalises canvassers and survey researchers alike.

The second dynamic is technological. Smartphones now integrate AI-powered call-filtering systems operating at three cumulative levels: at network level (operators' anti-spam services), at device level (native iOS and Android functions), and at the application layer (communitybased platforms such as Truecaller or Hiya). These systems, designed to block abusive canvassing, are blind to the purpose of calls: survey fieldwork, which structurally gene ates high call volumes and substantial non-pick-up rates, triggers exactly the same detection algorithms as commercial call centres.

The third dynamic is regulatory. The ARCEP's introduction of Verified Multipurpose Numbers (NPV) on 1 January 2023, intended to authenticate legitimate calls, paradoxically worsened the situation for researchers. By making su vey institutes explicitly identifiable through their number prefixes, this measure accelerated their conflation with commercial canvassing in users' minds. INED data show a first-call pick-up rate of 5.7% for NPV numbers versus 21.7% for standard numbers on landlines — a fourfold disadvantage. Moreover, operators regularly block survey institutes' NPV numbers mid-session without legal basis, forcing providers into a permanent rotation of their number pool — precisely the same strategy used by fraudulent actors that the regulation was designed to combat.

In response to this paradox, the Act of 30 June 2025 (loi n° 2025-594) created a new category of numbers — NPAIG (Numéros Polyvalents pour les Appels et messages d'Intérêt Général, i.e. Multipurpose Numbers for Calls and Messages of General Interest) — reserved for entities designated by ministerial order. While this legislative advance is real, its ope ational applicability remains suspended, at the time of this communication, pending publication of that order. It also faces a structural limitation: private filtering systems (scoring platforms, device manufacturers, operators) have no legal obligation to align their algorithms with national regulatory categories.

In addressing these challenges, two pitfalls must be avoided. First, assuming that contactability will stabilise spontaneously, when all available signals point to continued deterioration. Second, yielding to the appeal of synthetic data as a general solution: while such approaches offer promising prospects for strengthening the representativeness of under-surveyed subpopulations, their conditions of statistical validity remain to be rigorously established, and hasty adoption risks substituting a semblance of an answer for a real problem.

The appropriate response is collective and political. It requires that public and private actors in statistics and audience measurement speak with one voice to national and European regulators, drawing notably on Article 24 of the European Media Freedom Act (EMFA, 2024), which explicitly recognises the public-interest nature of independent audience measurement. Only by transcending institutional boundaries to collectively defend the infrastructure for producing representative data can the contactability crisis be met with a response commensurate with its democratic stakes.