Between November 2024 and June 2026, the reach on which a Tunisian media plan rested lost nearly one-third of its volume. This is neither a local anomaly, nor a measurement error, nor a temporary downturn: it is a reconfiguration of how the audience arrives — or no longer arrives — at news sites.
This study measures the scale of the phenomenon in Tunisia, compares it with the global shock observed in search traffic, reconstructs its chronology, and draws conclusions for those who buy, sell, or produce audience. We distinguish everywhere what the data demonstrate from what they suggest: this is the condition of a reading that holds.
1. The finding: −35 % in one year, of the same order as the global shock
Across a fixed basket of nine information sites tracked continuously, with scope kept constant, the cumulative audience falls from 59.3 million visits in the first half of 2025 to 38.3 million in the first half of 2026 — a decline of 35.5 %. The monthly series tells the same story: −33 % between November 2024 and October 2025, a stabilization on a low plateau at the start of 2026, then another drop in the second quarter.
Cumulative audience of the nine sites followed
Comparison of the first halves of 2025 and 2026 · In millions of visits
−35.5 %
of cumulative audience between the first half of 2025 and the first half of 2026.
What the monthly series also shows
−33 % between November 2024 and October 2025.
Stabilization on a low plateau at the beginning of 2026.
New decline in the second quarter of 2026.
Source: Tunisia Survey Observatory / Mindbox · Fixed basket of nine information sites.
This figure only makes sense when weighed against the global benchmark. According to the Reuters Institute, which uses data from analyst Chartbeat, traffic from Google Search to more than 2,500 information sites fell by one-third worldwide between November 2024 and November 2025, and by 38% in the United States. In the same window, Google Discover declined by 21%, and publishers anticipate a nearly twofold reduction in search referents within three years. Referents from social networks, for their part, fell by more than 40% over three years, for both Facebook and X.
Two independent instruments — one global, the other Tunisian — point to declines of similar magnitude, at the same time. This convergence does not by itself prove a cause-and-effect link, but it strongly reinforces the hypothesis of a common phenomenon: Tunisia is experiencing a decline of the same order of magnitude as the global shock to search. The Reuters Institute itself remains cautious about attribution: it notes the decline in search traffic, but stresses that it is not possible to determine precisely what share is due to AI-generated responses, with hard news queries largely outside the scope of AI Overviews.
The global benchmark of audience decline
Data Reuters Institute / Chartbeat · International comparisons useful to gauge the Tunisian market
Declines observed
Google Search – World
−33 %
Between November 2024 and November 2025
Google Search – United States
−38 %
Between November 2024 and November 2025
Google Discover
−21 %
Between November 2024 and November 2025
Social referents (Facebook, X)
> −40 %
Decline over three years
Publishers’ anticipation
Search referents in three years
≈ −50 %
Almost a division by two anticipated by publishers surveyed
Reading: the observed decline in search traffic is already pronounced at the global scale. Google Discover is retreating as well, while social referents have been declining even longer. Publishers now anticipate a continuation of the search retreat over the next three years.
Source: Reuters Institute for the Study of Journalism / Chartbeat. The periods are not all identical: Google Search and Discover are presented over one year, social referents over three years, and the nearly twofold reduction corresponds to publishers’ expectations, not to a drop already observed.
A precise note, because it conditions every reading: the global benchmark measures only search traffic, while our Tunisian figure covers total traffic. Tunisia is therefore experiencing a shock of the same magnitude at the same time, not necessarily the same shock in the strict sense. The future breakdown of traffic sources will determine what portion of this Tunisian decline genuinely stems from search, social, direct, or other channels.
2. Why these figures hold
A screenshot from an audience tool is not a study. Our estimates rest on a series covering twenty months, with three monthly readings unavailable, and above all on a verification rarely possible: for a site whose first-party analytics we know, external estimation matches the actual measurement to within 2 to 3 %, over six consecutive months.
This concordance has formal value only for this control site; it nonetheless gives solid confidence in the overall reading, particularly for high-volume titles, where estimates are the most reliable. It is the difference between observing a market and being able to respond to it.
3. It is not a matter of language, it is a matter of model
