This week on Media Finance Monitor
Protect the information ecosystem, not every organisation
AI might actually need the people it’s replacing
A partners only Capital Stacks session on the 3rd of November
How subscriptions came to fund half of a nonprofit local newsroom
22 active calls (4 new)
Media Finance Monitor is where newsroom leaders, funders, and media entrepreneurs come to understand how money moves through the information ecosystem. We decode funding, business models, policy, and platform shifts and turn them into strategies you can actually use.
Paid subscribers get premium editions, full access to the archive and all our research reports, for about the cost of a pistachio éclair and a good espresso a month.
For deeper engagement, our Partnership Program offers one-on-one strategic advisory, curated briefings, and access to our internal databases and invite-only events.
Protect the information ecosystem, not every organisation
(by Peter)
Years ago, a media development organisation asked me for the home addresses and phone numbers of two emergency contacts for a mostly online consultancy assignment. I am not too precious about my data, but I couldn’t understand why my wife’s mother’s maiden name or her personal mobile number was necessary for me to run a service design exercise with a small investigative newsroom. I politely refused.
I have always thought this slightly Kafkaesque encounter illustrates much of what is wrong with this kind of development work: bloated bureaucracies, institutional inertia, excessive safetyism, and an obsession with process that sometimes leaves little room to consider whether any of it improves the work.
The 2025 funding crunch has cost good people their jobs and ended valuable work. Organisations have lost expertise built over years, and communities have lost support they will not easily find elsewhere. I know some of the people affected and there is no clever argument about efficiency that makes those losses less painful. But when I think about that emergency contact form and what it represented to me, maybe some change is actually welcome.
Take Internews, whose 2024 impact update described a global community of more than 1,100. With legal entities in the US, UK and France, it ran large, often government-funded projects around the world.
The wider Internews network has endured a brutal contraction following the collapse of US foreign aid. Staff, partnerships and country offices were lost. I knew people who got laid off, we actually hired some of them.
Now Internews Europe and Internews International are becoming Informed. US-based Internews retains its name; the organisations are moving towards more independent operating models.
I like the name. Public interest information increasingly feels like the relevant building block, whether delivered by a newsroom, a creator or something else. The name puts the emphasis where it belongs. And while it’s probably too early to judge the results of the restructuring, I find the willingness to rethink the organisation encouraging.
Or take the Plus alliance. In June, the European Journalism Centre joined Code for Africa, the International Center for Journalists and PROTO, which established the alliance in 2024. Members remain distinct entities, while pooling operations, infrastructure and resources. This is substantial operational consolidation, even without everyone disappearing into one legal entity.
This is very welcome. Organisations recognising that they might serve their constituents better together deserve credit. This does not negate the human and professional cost of consolidation, but preserving the work is not the same as preserving every organisation intact.
We have been advising an international newsroom that is exploring a merger. It produces work I value in a funding environment that has become punishing. I admire their willingness to confront the budget and consider surrendering autonomy in order to preserve the mission.
Funders should help organisations explore consolidation, especially in publishing. Support should make a voluntary decision possible; it should not become a donor instructing two dependent grantees to marry. There are precedents in the US and in Europe, V-Ventures is getting a CFO-as-a-service experiment off the ground.
Consolidation can destroy useful capacity, but it can also preserve more of it than several organisations struggling separately until their money runs out.
AI might actually need the people it’s replacing
(by Peter)
Time to look in on the always heartwarming world of AI-mediated information and what it does to jobs in publishing.
Future, the British publisher of Tom’s Guide, TechRadar and Android Central, is cutting jobs for the second time this year. Press Gazette reports that Tom’s Guide has lost 55% of its UK audience since the month before Google’s AI Overviews launched there, TechRadar 35%. The Verge reports that everyone at Android Central has been laid off; Future says the site will continue. In January the company made 45 editorial redundancies across its tech titles.
While I really like reading them, I’m ready to concede that gadget reviews are not the most essential form of public interest information. Still, they are a useful case-study, because product advice is exactly what people now ask a chatbot for.
I recently got a MacBook Neo for a family member. I researched it on the old-fashioned internet, reading reviews and comparing benchmarks. We pay for the Wirecutter, which recommends it too.
I could probably have asked a chatbot and reached the same conclusion more conveniently. But my browsing generated something useful for publishers: visits against which they could sell advertising, and opportunities to earn affiliate commissions. The chatbot might deliver the recommendation without delivering any value for people who created the content.
The structural problem here is that the chatbot can only answer reliably, because somebody actually tested the laptop. Apple says the battery lasts 16 hours. Tom’s Guide ran one down and got 13 hours and 28 minutes, the Wirecutter had it at 14 hours. A language model cannot do that. I don’t think most readers really care who does the testing, but they do care whether the numbers are right. Once the answers stop being right, because it runs on marketing claims rather than verified information, convenience will not save the product.
There are some attempts exploring payment. Google is reportedly piloting compensation linked to content’s contribution to AI answers, and Perplexity has announced revenue sharing. But as of today, no one has a dependable, widely accessible replacement for the revenue being lost.
The most serious industry response so far is SPUR, a coalition formed in March that includes the BBC, the Financial Times, the Guardian, Sky, the Telegraph, Mediahuis, the AP and others. On 2 October it published a content telemetry standard and invited OpenAI, Anthropic, Google, Meta and Microsoft onto an advisory board.
(I grew up watching Star Trek, so I welcome an industry in which “telemetry standards” are part of my working vocabulary. The underlying economics are slightly less glorious.)
Think of the standard as a receipt. It follows an article through five stages: retrieved, grounded (fed to the model while it writes the answer), cited, presented on screen and engaged with. “Grounded but not cited“ means your work shaped the answer and you got no credit. The standard sets no prices, but supplies the numbers a deal could be built on.
All very reasonable, but it only works if the AI companies turn up. Asked whether it had accepted the invitation, Google said: “We frequently engage with SPUR and other associations on a variety of topics.“ while the others didn’t answer. Large publishers acting together is the only version of this with any leverage; alone, none of them has enough. I like the free market (I own several T-shirts saying so), but I am less sure how long we can wait for it to sort this out before some policy leverage becomes necessary.
The models can answer today because people were paid to find things out. Once they have eaten all of that (and put enough people out of work), the poor things starve.
A partners only Capital Stacks session on the 3rd of November (Save the date and/or join the Partnership Program!)
(by Peter)
We’ll do a Partner’s only, closed door, exclusive, Chatham House, VIP online presentation of the Capital Stacks report and we’ll answer any questions you may have about the findings on the 3rd of November from 15:00 CET. Organizations and individuals already in the Partnership Program will receive their invitations next week, and if you are not a CSM Partner yet, this is a good time to reconsider. There are many valuable benefits (really), you can read about them here.
How subscriptions came to fund half of a nonprofit local newsroom
(by Maria)
Sixteen thousand paying readers generate about $3 million a year for the Daily Memphian, enough to cover roughly half of the nonprofit outlet’s $6 million annual operation. Launched in 2018 to serve Memphis, with $7 million in philanthropic backing, it now combines subscriptions with advertising, events and philanthropy. In a recent interview with Medill’s Local News Initiative, co-founder and former CEO Eric Barnes laid out the economics and operating choices behind that model.
Recurring reader revenue can make income less vulnerable to individual losses. Thousands of individual subscribers account for half of total revenue, spreading that income across many relatively small payments rather than a smaller number of large accounts. An annual subscription costs $178, so losing one subscriber has a much smaller impact than losing, for example, a $20,000 advertiser or a large grant. The outlet also favors longer commitments: around 65–70% of subscribers pay annually, and it A/B tests offers to encourage more readers to choose annual plans.
Revenue generation needs dedicated capacity too. Alongside around 40 full-time newsroom staff, roughly 12 people work across advertising, technology, marketing, business management and customer service. Customer service is handled by staff in the same Memphis office, keeping feedback from paying readers stays close to the organization. On the advertising side, 99% of sales are handled in the local market by the outlet’s own representatives instead of through programmatic networks. In both cases, the relationships with subscribers and advertisers remain relatively direct.
A paywall does not have to mean all-or-nothing access. Beyond its paid base, the outlet has 50,000–60,000 free registered users who can read four articles a month without paying, while public schools, libraries, senior living facilities and affordable housing receive free access. Its newsletters reach around 160,000 email subscribers, while hundreds of thousands more follow its social channels, where headlines and short updates provide another way to keep up with local news without reading every full article. Keeping some content freely available is part of the nonprofit’s mission; Barnes said a for-profit version would probably charge more and give away less.
Audience acquisition can extend well beyond platform referrals. Direct visits to the site and app are now its largest traffic source, with email close behind and Google third. The outlet has long restricted how much of its content Google and Facebook can access, even at the expense of its search ranking, while also using local channels such as events, billboards, TV advertising and bus wraps. As referrals from search and social weaken, it is putting more emphasis on email and this “real-world marketing.” So far, those efforts have helped mitigate traffic losses, even if they have not driven a major increase in traffic.
This piece is part of a series focusing on local and community journalism and is supported by the LimeNet project and the European Union.
Here are the 22 active calls (4 new), with the largest at the top: