Newsletter

Are we measuring the wrong thing? What Reuters’ Digital News Report tells us about engagement

The old metrics of clicks, views and app opens no longer tell publishers what really matters. Using insights from the Reuters Institute’s Digital News Report 2026, we make the case for measuring deeper engagement in owned products.

23rd July 2026

As trust declines and audience behaviour shifts, publishers should stop relying on legacy traffic metrics and start measuring the deeper engagement that drives loyalty, habit and revenue.

Publishers still tend to measure value using the same metrics they relied on a decade ago, when the priority was maximising views, driving clicks and fine-tuning the traffic mix. But those measures were built for a very different media environment, and the Reuters Institute’s Digital News Report 2026 suggests they now reward behaviours that increasingly conflict with trust, comprehension and the durable audience relationships publishers need to sustain their products.

In this week’s Media Bulletin, we’ll look at what the recently released Reuters report says about platforms, AI and audience behaviour. We’ll explore what good engagement should mean inside apps, websites and other owned environments and how we believe publishers should measure it going forward.

Engagement in the era of AI

In an environment where platforms and AI increasingly mediate how people find news, simple reach metrics like clicks, page views and app opens can become misleading. Traffic from channels such as Google Discover and emerging AI chatbots produces eye‑catching spikes in visits, but much of that attention is shallow, volatile and skewed towards lower‑value content that does little to grow subscriptions or long‑term loyalty. The same pattern shows up in push campaigns built purely around maximising open rates, where aggressive headline optimisation lifts the number of opens but leaves subsequent engagement on the article lacking.

The Digital News Report 2026 puts these changes in a wider industry context. Globally, trust in news has fallen to around 37%, while roughly four in ten people say they sometimes or often avoid the news. At the same time, social and video platforms have overtaken TV and publishers’ own websites and apps as the most widely used sources of news, shifting the centre of gravity decisively towards third‑party feeds. More of the audience’s attention now sits in environments that amplify volume and velocity but do not necessarily reward depth, trust or direct relationships.

Many of the metrics that still dominate newsroom and product dashboards were built for a world where platform traffic looked like free growth. In 2026, those same metrics can incentivise behaviours that run directly against what fatigued, distrustful audiences say they want. Publishers now need to pivot towards measures of deeper engagement across formats and journeys. This includes time spent with high‑quality articles, completion rates, repeat visits to apps and sites, and sustained use of richer formats like audio and video. Optimising for these kinds of signals is what will build durable subscription businesses and resilient audience relationships in an AI‑disrupted ecosystem.

Platformisation and the drift away from direct news

The Reuters report describes an “unsettled” information environment in which social media, video platforms and other intermediaries have overtaken news organisations’ own sites and apps as the primary way people access news globally. These environments are engineered to maximise attention, speed and emotional response, not to reward careful reporting, nuance or long‑term relationships with a single brand. As audiences spend more of their time in feeds that optimise for raw attention, publishers risk importing those same incentives into their own products and metrics, chasing volume and volatility instead of durable value.

Concern about misinformation and “AI slop” is high and only a minority say they trust news in social feeds or AI answers. People continue to place significantly higher trust in the brands they choose deliberately than in generic platform environments. That means that publishers can either compete on platforms’ terms, optimising for the same shallow engagement signals, or differentiate by doubling down on quality, transparency and editorial standards in their own destinations, where trust and membership can actually be built.

AI makes that choice more urgent. Around one in ten people globally now use AI chatbots for news each week, with usage skewed towards younger, heavy‑news users who frequently ask follow‑up questions and seek interpretation rather than just headlines. In many of these AI‑mediated journeys, the most meaningful interaction with a publisher’s journalism may happen without a click‑through at all. If newsrooms and product teams rely on KPIs inherited from attention platforms, they will systematically under‑value those deeper, trust‑building encounters.

The trouble with optimising for attention

With the centre of news consumption moving into attention‑optimised environments, there are three primary reasons why only looking at legacy metrics is not the best approach.

First, those metrics are structurally biased towards platform behaviour. Click‑through rates, impressions and views were borrowed from environments where content competes in a mixed feed and the easiest levers are speed, emotion and constant posting. Those indicators are very good at telling you how you perform inside someone else’s attention system, but very bad at telling you whether you are building a distinctive, high‑quality product and a resilient audience of your own.

Second, they conflate attention with understanding and value. A second spent glancing at an article’s headline can look identical to 5 minutes spent reading a carefully constructed explainer or listening to an podcast. The report’s data on falling interest in news and widespread news avoidance suggest that audiences are increasingly overwhelmed and fatigued, which means that “attention spikes” can sit alongside a growing desire to step away rather than go deeper, precisely the opposite of what quality journalism needs.

Third, they ignore or under‑count multi‑step journeys. AI chatbots, creator channels and algorithmic video feeds are now central to how many people encounter and work with news. Much of the trust‑building and habit formation happens in these layers, but traditional referrer‑based metrics see only the occasional click‑through and treat dashboards full of on‑site spikes as a proxy for success.

Rethinking “what matters” inside owned products

If traditional engagement metrics are misaligned, then publishers need a different approach for measuring good engagement in their apps and websites. In these environments, they have far more control over the product and far better signals, so the bar should be higher.

Depth should become the primary signal. Instead of assuming that a click is the end of the story, publishers should look at how often people complete an edition or timeline, reach meaningful scroll‑depth bands on key articles or finish listening to an audio version in the app. These behaviours map much more closely to the “did this help me understand what is going on and why it matters?” test the Digital News Report emphasises and they are far less vulnerable to the distortions created by outrage and sensationalism.

More attention should be paid to habit than traffic spikes. Registration and app data make it possible to track session frequency, recency and to distinguish “super users” from more passive audiences. With organic search traffic under pressure and platform algorithms increasingly volatile, the Reuters report argues for focusing on owned pathways such as apps, email and direct visits and measuring how well those routes convert and retain users over time. Elevating metrics like the share of sessions from owned channels and registration‑to‑subscriber conversion by channel makes it easier to see whether a product is accumulating value or simply gaining attention.

Measuring AI without repeating old mistakes

Publishers also need to consider their approach to measuring the impact of visibility in AI platforms. The temptation is to bolt on a new layer of metrics, such as appearances in chatbot answers and declare success when those numbers rise. Given the experience of the past decade and the fact that news found via social media and AI answers is trusted significantly less than news from established brands, treating AI visibility as a vanity KPI would repeat familiar mistakes.

The Digital News Report’s AI chapter shows that usage and trust are tightly connected. People who choose to use AI chatbots for news are much more likely to trust them than those who do not and usage is still concentrated among power users who rely on AI for interpretation and making stories easier to understand. If publishers simply optimise for whatever performs best inside AI interfaces, they risk rebuilding the same attention‑driven trap in a new medium.

A better approach is to treat AI‑generated sessions as a distinct segment wherever possible and to compare their depth, habit and conversion behaviour with other acquisition routes rather than just counting impressions. AI traffic metrics should remain secondary unless they demonstrably connect to meaningful product outcomes. AI tools themselves can be used to improve measurement, for example, by classifying content to study completion, summarising open‑text feedback at scale, or helping teams see which formats and topics are most associated with sustained use and lower avoidance.

The metrics that filled publisher dashboards over the past decade are no longer a reliable proxy for the relationships a sustainable news business needs. The opportunity now is to make those relationships visible inside apps and other owned environments, and to let those signals, rather than platform‑style attention numbers, guide decisions about formats, paywalls and the role of AI in products.