Editor’s Pick Products & Formats Editors Picks

Sky News has a platform deal. The audience is the harder part

Ivan Massow 3 Nov 2026 · 5 min read

A visitor is swept through a revolving door and straight back outside while a red-clad host gestures proudly.

Cartoon · AI-generated

Sky News has agreed on the technology for its next app and website. Winning a place in people’s daily lives is a less settled matter. The expanded Arc XP partnership promises video, personalisation and room for paid products. None of those capabilities, on its own, explains why a reader should make another visit.

That is the commercial question behind the announcement. A better service could save readers time and give Sky more chances to earn from its reporting. The newsroom and product teams must supply the continuing work. Before celebrating the new platform, publishers should ask what it will make possible that audiences will value enough to seek out.

A deal is the beginning of the test

Announced on 2 November, the agreement targets a new Sky News app and website by the end of 2027. It extends a collaboration announced in November 2025, which included work on conversational search. The change is a firmer product commitment, rather than evidence that the audience strategy has already worked.

Advertisement · Noah

A content management system, or CMS, organises and publishes a newsroom’s work. The interesting part of Sky’s announcement is the attempt to connect that machinery to a reader’s experience: finding a relevant report, following a story and perhaps buying something more specialised. A procurement announcement can describe the means. It cannot establish the demand.

Sky says video views in its app have more than doubled year on year. The release supplies no absolute baseline, so the number cannot tell readers how large that activity is. Nor does it show whether more people are watching, existing viewers are watching more, or either group is becoming more valuable commercially.

The distinction matters because a publisher can improve a consumption figure without solving its business problem. A useful test would examine whether people return after their first visit, what brings them back and how that behaviour supports advertising or payment. Launching a new interface is a milestone; those are the outcomes that would justify it.

The specialist offer has a clearer proposition

Sky’s Defence Insider offers something more specific to examine. Launched on 10 September at £14.99 a month, the standalone app combines specialist reporting and analysis with opportunities to question journalists and experts. It sits alongside the free general news app.

The proposition is access to expertise on a subject that matters to the buyer. Video is the delivery format. A reader might value an explanation from Deborah Haynes or Michael Clarke enough to pay; the same reader does not necessarily want a subscription to a more polished version of general news.

The launch announcement gives no subscriber or retention figures. The expanded Arc agreement also stops short of announcing a general Sky News paywall. The evidence establishes a specialist product and a broader commercial ambition, with the relationship between them still to be demonstrated.

That makes Defence Insider a useful test of Sky’s strategy. If people pay repeatedly for distinctive reporting and expert access, there is a reason to develop the service. If the attraction fades after the initial interest, another set of technical features will be an expensive response to a weak proposition.

Personalisation has to save the reader effort

Personalisation can sound more ambitious than the useful task it performs. Remembering where a reader left off, showing what changed and avoiding a page full of material already read could be valuable. These are intelligible benefits that can be tested without promising to reinvent news.

Arc’s September launch of Compass describes catch-up features, connected story timelines and controls for editors to promote or demote material. The vendor says the tool works with other content management systems. Publishers considering it can therefore ask whether a limited experiment would answer their question before committing to a complete migration. Sky has not confirmed that Compass is part of its new products.

Consider a hypothetical reader returning to a local planning dispute on Friday after reading about it on Monday. The useful service is the decision, the reasons and the next meeting. Recognition software is only part of that service. Someone must connect the reports, keep them current and decide which change deserves attention.

Germany’s Madsack supplies a concrete comparison. A Google Cloud case study describes bringing data together across regional titles and creating a weekly email of popular stories each subscriber had not read. The account comes from a supplier and does not independently establish a retention gain. It does, however, describe a service a reader can understand, rather than an abstract promise of engagement.

New formats bring new editorial liabilities

Conversational search adds another way into a publisher’s reporting. It also adds questions about what happens when reports change, sources disagree or the answer is missing. Readers need to reach the underlying work and understand its date. Those requirements should be part of the product specification, with people responsible for correcting failures.

The Washington Post’s Your Personal Podcast provides a caution about turning existing journalism into a personalised output. Semafor reported accuracy and standards problems in December 2025 after examining internal tests. The Post’s current help page still describes episodes generated from its articles, but provides no basis for concluding that those earlier problems have been resolved.

That example does not establish a fault in Sky’s plans, and generating narration is different from retrieving an existing report. It does show why a publisher should test the experience it actually intends to release. The editorial standard cannot rest on the assumption that material drawn from a trusted archive will remain accurate after it has been transformed.

The case for AI here is practical. A tool that helps readers find and understand good reporting can be useful. Its value should survive questions about errors, maintenance and cost. Treating those questions as resistance to innovation would give the supplier an easier time than the publisher’s own newsroom.

Shared software leaves the strategy with the publisher

An analysis of publishing coverage by Noah, a research and forecasting service, identified a similar ambition at Editoria Italia. Its February agreement with Arc covers brands including Libero, Il Giornale and Il Tempo. The stated aim is to simplify operations and support experimentation across titles. Those are aims, not published evidence of improved earnings.

Shared machinery may spare each title from rebuilding the same capabilities. It also creates a dependency. A publisher considering a migration should establish how reporting, video, account records and permissions can be exported, what support it is buying and how much work remains with its own teams.

A catch-up service needs reliable updates. Expert Q&As need journalists’ time. Video needs captions and an experience that works without sound. These continuing costs belong in the business case, alongside the software and launch budget. Moving them out of the procurement presentation does not make them disappear.

Sky has named a supplier and a timetable. The harder evidence will come from the audience: regular use, a reason to choose its journalism and a willingness to pay where there is a paid offer. A new platform may help produce those results. It should be judged on them, rather than congratulated in advance.

Research note

Research combined an analysis of publishing coverage with checks of company announcements, product information and original reporting. Automated forecasts were unavailable. No replacement probabilities have been supplied.

Sources

Source: Tomorrow’s Publisher