A month ago in this space, the destination died. Then scarcity shifted from argument to measurement. Then the traffic bargain got its obituary on three continents at once. Last week, we counted the assets publishers actually own: trust, habits, and the inbox.

This week, the graders showed up.

Three signals, all within 48 hours of each other:

• Punished for the curve: The New York Times Company reported a quarter most publishers would frame and hang on the wall — revenue up 11.2% to US$762.5 million, 280,000 net digital subscribers added, digital advertising up 20.7%, earnings ahead of consensus. The stock fell almost 9% before the market opened because management guided third-quarter digital subscription revenue growth down to 9%-11%. The best consumer subscription business in news beat expectations on nearly every line and was punished for the shape of its curve.

• Rewarded for the exit: The same day, Thomson Reuters reported 9% revenue growth, raised full-year guidance, talked up “fiduciary-grade AI” for professional markets — and sold a majority stake in its global print business to KKR for roughly US$500 million. The market applauded.

• Confirmed in the org chart: Axel Springer made Christian Baesler the permanent CEO of Business Insider with a mandate that reads like a summary of this blog’s July editions: four subscription verticals, 10 million YouTube subscribers, a streaming app, a FAST channel — and no plan that depends on Web traffic.

One flagship punished, one rewarded, one restructured — all for the same underlying reason. Capital markets have stopped grading news companies on what they earned last quarter and started grading them on where they are going.

Here is the storyline I take from the past five days of the INMA conversation: Three different graders are now scoring this industry’s transition.

1. Markets grade quarterly.
2. Platforms grade daily.
3. Readers grade for life.

Only one of those grades compounds — and this week produced an unusual amount of evidence about how to earn it.

Africa turned in the homework

Last week I called our Africa Newsrooms & Tech Summit the ownership playbook on hard mode. This week, the case studies landed on the INMA Conference Blog — and hard mode looks less like an exception and more like a preview of everyone’s next level.

What strikes me reading them together: Not one of these stories is about scale. Every one is about discipline.

• Doing less, converting more: Nation Media Group in Kenya cut daily story output nearly in half — and got higher traffic, stronger subscription conversion, and less burnout. Harry Misiko moved the newsroom’s daily question from “Can we cover this?” to “Should we even cover this?”— scoring every story against editorial value, audience need, business value, and feasibility.

• One workflow, not five silos: Uganda’s Vision Group collapsed five platforms into a single workflow, rebuilt around content hubs feeding digital first, and grew its ViDE super-app from 2,000 users to 10 million downloads and roughly five million actives. Penlope Nankunda’s framing belongs on the wall of every transformation office: “Sustainability is not a destination. It’s continuous adaptation.”

• Membership as a nine-year boulder: South Africa’s Daily Maverick built Maverick Insider to nearly 50% of organisational revenue — without a paywall, over nine uphill years. Julia Harris shared the operational detail that should make every reader-revenue leader sit up. Roughly one in seven members had failed payments, making payment recovery a major revenue lever. And she explained why they built their own tools rather than buying: “The thing that drives your revenue is proprietary to your model.”

• AI with a written constitution: Nigeria’s Punch embedded AI across the newsroom under a formal ethics policy — AI proofreads, suggests headlines, checks style; it does not write stories, and humans make every final call. BusinessDay said the same thing in fewer words: “We use AI as an assistant and not as a reporter.”

• Fail fast, rebuild on purpose: Briefly News offered the candid failure-first arc. Its off-the-shelf custom GPT hallucinated errors and wasted writers’ time, so the team rebuilt a purpose-built, workflow-embedded tool suite — and watched one representative writer’s weekly performance jump more than 150%.

If I had to pin one quote from the Africa summit above the desk, it is Punch Editor Oyetunji Abioye’s: “The winning newsroom will not necessarily be the one using the most AI. It will be the one using AI most responsibly.”

That is not a compliance statement. It is a competitive strategy.

The business model learns to travel

The July 8 edition of this blog argued that formats must travel to readers. This week, the INMA conversation pushed that logic one step further: The business model has to travel, too.

• Subscriptions that live with the subscriber: Jodie Hopperton’s Product & Tech Initiative blog asked the question out loud — if distribution is happening on platforms, shouldn't subscriptions work there, too? Picture an FT subscriber authenticating inside ChatGPT to unlock premium reporting or a local news subscriber receiving gated journalism over WhatsApp. The precedents half-exist — Apple Podcasts subscriptions, Spotify’s Open Access — and the blocker, Jodie argues, is prioritisation, not technology. Her prescription is the right one: Stop negotiating with AI companies solely for licensing checks, and start collectively pushing for an open authentication layer.

• Format first, brand second: Her second post this week explains the urgency. Overall appetite for video is booming, yet demand for news video on publisher Web sites is declining — because we keep yanking people out of clean, personalised feeds into cluttered homepages. Her answer is “transition products” that meet the arriving user in the format they just left.

• Creators are grading our homework: Kerstin Hasse, lead of the INMA Young Audiences Initiative, supplied the uncomfortable mirror — content creators execute the user needs model better than the newsrooms that invented it. With 27% of people globally now getting news from creators weekly, the root cause she identifies is not talent but measurement: Newsrooms reward volume, which structurally favours commodity updates over the formats that build loyalty. That diagnosis rhymes exactly with Nation Media Group’s output cut. The metric is the strategy.

And look how many of this week’s external headlines are the same idea wearing different clothes:

• The Economist launched a lower-priced tier built on video, audio, games, and podcasts for audiences who will never start with the weekly edition — vertical video views up 130% year over year.

• The New York Times is putting Wordle on NBC prime time explicitly to reactivate lapsed players, with plays down to 4.2 billion from 5.3 billion — broadcast television as top-of-funnel for a subscription games bundle.

• Podcasters are colonising free ad-supported TV channels — Scott Galloway, Crooked Media, The Ringer — because their content was already video and the living room was just sitting there.

• Even entertainment tells the story: Southeast Asia’s streaming boom shows attention traveling globally — and the money following the ones who travel with it.

The destination-era instinct was to pull everything home. The emerging instinct is to send the product — and the payment relationship — wherever the audience already is.

Grading without a referee

If markets are one grader and readers another, the platforms appointed themselves a third this week — and they are writing the rubric as they go:

• Google put a price on participation. Press Gazette reported Google is now embedding full Top Stories carousels inside AI Overviews — appearing in 15.5% of trending news queries globally, 17.5% in the UK. The catch: Publishers that use opt-out controls to keep their content away from AI features may lose eligibility for the placement. Visibility inside the AI answer is the carrot; opting out now carries a measurable traffic cost. That’s not a partnership. That’s a toll booth.

• The ad stack is arriving before the rules do. Gabe Dorosz, INMA Advertising Initiative lead, flagged on INMA Slack this week that OpenAI is integrating with AppsFlyer to bring in-app conversion measurement to ChatGPT ads — GrubHub and 40+ brands are testing, and GrubHub has already moved from CPM to conversion-based bidding. Note what GrubHub told Adweek: Non-branded, discovery-style search is shifting to ChatGPT. Measurement gaps were the only thing holding spend back — and OpenAI, chasing US$2.3 billion in ad revenue this year and US$100 billion by 2030, is closing them one ad tech partnership at a time. Emarketer’s Nate Elliot notes Google certified 1,100+ ad tech partners over 20 years; OpenAI can reach parity eventually, “but it’ll take time ... time seems to be the one thing they don't have.” Translation for publishers is the AI answer layer is becoming a performance ad channel with real attribution — while news content inside it still has neither.

• A manipulation industry is already running. The Verge chronicled the gold rush of firms promising brands visibility inside ChatGPT, Gemini, and AI search results — a booming cottage industry veteran search experts describe as somewhere between repackaged SEO and outright snake oil, complete with tactics designed to game what the models say.

• We cannot even referee authorship. Spain’s Laboratorio de Periodismo surveyed the state of AI detection and concluded no current technology can certify a human wrote something — one leading detector, claiming 99.98% accuracy, was fooled 86% of the time in adversarial testing. “Human-made” as a trust signal will have to be an editorial process, not a product feature.

• The old referees are leaving the field. X settled its long-running suit against the World Federation of Advertisers, and buried in the settlement is the consequential clause that the WFA commits never to restart GARM or any similar brand-safety initiative. Collective standard-setting by advertisers is over — litigation as leverage worked.

• And a new referee steps in. Australia finalised its News Bargaining Incentive: a levy of up to 2.5% of Australian digital ad revenue on large platforms, waived for those that strike deals with news publishers. The world’s most-watched experiment in taxing platforms into paying for news is now a template other governments will study.

The pattern across all six stories is that the rules — and now the revenue infrastructure — of the new distribution layer are being written by platforms, vendors, courts, and governments. Everyone except publishers. Which is precisely why Jodie’s call for a collective, publisher-defined authentication standard deserves more attention than it will probably get.

The grade that compounds

So what does the reader’s report card look like? This week supplied more hard numbers than any week in recent memory:

• The FT put numbers on belonging. Hannah Sarney shared the economics of community: commenters are 4x more engaged than non-commenters, monthly subscribers who comment retain 24% better, and trial subscribers who comment are 76% more likely to convert to paid. Yet only about 4% of FT subscribers comment while more than half read comments — which makes the FT’s pairing of human editors with AI moderation and AI-drafted discussion prompts not a nice-to-have but an activation strategy for one of the largest untapped conversion levers in the building.

• Local publishers are running streaming’s playbook. The American Press Institute rounded up subscription experiments: The Boston Globe converting household password-sharing into family plans, The Texas Tribune bundling festival access for donors, and The Minnesota Star Tribune’s high school sports vertical hitting a 5% conversion rate on a single livestream — triple its previous attempts.

• Usefulness still moves the needle. Nieman Lab’s quarterly traffic reviews showed what reliably earns local attention: The Philadelphia Inquirer engineering a 17% lift with a deliberate weekend lifestyle franchise, The Orange County Register spiking 54% on live safety utility during a chemical emergency, and public media growing through accountability collaborations and election guides.

• South Asia priced the trust premium. In Gabe’s post on advertising in South Asia, where ad markets grow at double the global rate, an Inteliphyle trust-attention-conversion index scores newspapers at 5.37 against 1.0 for online news — and former Tata Motors CMO Shubhranshu Singh compressed the entire thesis into eight words: “Reach is a commodity. Trust is the premium.”

• Latin America is already operating the successor model. In Gabe’s companion post, the diversified ad business is not a strategy deck — it is running at scale: Editora Globo earning 60% of ad revenue from integrated projects and nearly one-third of total revenue from events; Grupo RBS operating a six-person journalist-creator video team at a 70% margin; Milenio’s loyalty members consuming 7x more content.

Readers, it turns out, are the only grader who pays twice — once with attention, once with money. And when they do, advertisers pay a third time for proximity to the relationship.

The bottom line

Braid the strands, and the storyline of the week reads like this: The market now grades direction, not results — ask The New York Times, which beat everything and fell 9%. The platforms grade compliance, and they set the toll — ask anyone weighing Google’s opt-out penalty or watching OpenAI wire conversion measurement into ChatGPT while publishers wait for an authentication layer.

But readers grade usefulness, habit, and belonging — and theirs is the only grade that compounds, quarter after quarter, the way Daily Maverick’s nine-year boulder-push and the FT’s commenter cohorts compound.

So:

• Ship the playbook, not the memo. Africa just showed the discipline it takes.

• Let the business model travel with the audience — from ChatGPT to WhatsApp to the living-room TV.

• Push collectively for standards on the AI layer before the toll booths and the ad stack finish going up without us.

• Measure what builds loyalty, because the metric is the strategy.

And study for the reader. It is the only exam with compounding returns.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .