SODP Dispatch - 19 February 2026

Readers Prefer To Click On A Clear, Simple Headline − Like This One

By David Markowitz

In an era when people trust news less than ever, how can journalists break through and attract the attention of average people to provide information about their communities, the nation and the world?

By not complicating things.

Our research, published in Science Advances, shows that simple headlines significantly increase article engagement and clicks compared with headlines that use complex language.

In our research, typical news readers preferred simple headlines over complex ones. But importantly, we found that those who actually write headlines – journalists themselves – did not.

We first used data from The Washington Post and Upworthy to see how language features, such as word length and how common a word is, changed how many people clicked on an article’s headline. These datasets included over 31,000 randomized experiments – also known as A/B tests – that compared two or more headline versions of the same underlying article to determine which one generated the most clicks.

Headlines with more common words – simple words like “job” instead of “occupation” – shorter headlines, and those communicated in a narrative style, with more pronouns compared with prepositions, received more clicks. For example, The Washington Post headline, “Meghan and Harry are talking to Oprah. Here’s why they shouldn’t say too much” outperformed the alternative headline, “Are Meghan and Harry spilling royal tea to Oprah? Don’t bet on it.” This example illustrates how sometimes a more straightforward headline can generate more interest.

In follow-up laboratory experiments, we found that typical news readers focused more on simple rather than complex headlines because the writing was easier to understand. When journalists participated in the same experiments, they didn’t show any preference for simple headlines over complex ones. Put differently: Those who write the news appear to be less responsive to simple writing than regular audiences.

The RPM vs. Traffic Tradeoff Poll

Would you trade 30% of your traffic for 2x your RPM?

• Yes—revenue per visitor

• No—traffic compounds over time

📊 Subscription LTV Calculator

Most subscription teams don't know their real LTV, they're working off blended averages and rough guesses. That's how profitable campaigns get paused and unprofitable ones get scaled. This free calculator tackles that gap by modelling what actually moves LTV: churn, renewal rates, annual plan adoption, and cohort behaviour over time. It outputs your net LTV per install, real Max CPA, and cohort accumulation, the same diagnostic framework used in live client work. If your team is making acquisition decisions without this visibility, it's worth a look.

Read more ▸

🤖 Metehan.ai

Most people interact with AI search without questioning what's happening under the hood, metehan.ai changes that. This breakdown reveals how tokenisation quietly shapes everything, from which languages get the most out of AI systems (English sits at 100% efficiency; Turkish prose at just 61%) to why AI hallucinates URLs (each token is generated independently with no ground truth to anchor it). It also unpacks why prose outperforms structured lists in tokenisation, why frontloading content matters, and why hidden links are the most token-efficient citation format. Even your HTML cleanliness has cost implications for how crawlers process your pages. In an era of soaring AI infrastructure costs, these aren't just technical quirks, they're biases baked into the system.

Read more ▸

By Vahe Arabian

🔍 Your Analytics Platform Wasn't Built for Publishing, Here's How to Fix That

Most publishers rely on external analytics platforms that were built for e-commerce and lead generation, not publishing. These tools measure traffic well but can't answer the questions that actually drive publishing businesses: Which content converts readers to subscribers? What behaviour predicts churn? Which authors drive retention? When your business model depends on understanding reader value beyond clicks, generic analytics consistently falls short.

The solution isn't abandoning external platforms, it's building a foundational data layer beneath them. Publishers who control their own analytics infrastructure can answer business-critical questions that generic tools weren't designed to address. External platforms still provide value for traffic monitoring and benchmark comparisons, but they measure what happened, not why readers converted or churned. That distinction matters enormously when editorial and commercial decisions depend on understanding reader behaviour at a deeper level.

Where to start : Identify your single most critical business question, whether subscription conversion or churn prevention, and build the minimum infrastructure to answer it. Practically, this means implementing server-side event tracking for paywall interactions, metered article counts, registration attempts, and subscription upgrades; establishing a unified reader identity framework that recognises readers across anonymous browsing, registration, and subscription states; and storing raw event data in a warehouse you control for flexible analysis and long-term access.

Privacy regulations make this even more urgent. Cookie restrictions and consent requirements mean publishers increasingly must manage how data is collected, stored, and processed. That responsibility can't be outsourced to a third-party platform indefinitely, and the publishers who act early will have a significant structural advantage over those who don't.

The payoff is significant. Custom infrastructure enables predictive churn modelling, dynamic paywall optimisation, personalised content recommendations, and automated audience segmentation, capabilities external platforms simply don't provide. Editorial teams gain clear visibility into what drives business outcomes, not just clicks. When external platforms change pricing or deprecate features, you're not held hostage. Your historical data remains accessible, and your business intelligence doesn't reset.

Start small, solve one problem well, and expand as your capabilities grow.

Key takeaways for publishers:

• External analytics weren't built for publisher-specific questions about conversion, churn, and content attribution

• Privacy regulations require publishers to control their own data collection infrastructure

• Start by identifying your most critical business question, then build minimum infrastructure to answer it

• Implement server-side event tracking for paywall interactions, registrations, and subscription behaviours

• Store raw event data in warehouses you control for flexible analysis and historical access

Today’s media landscape requires focus on AI, diversified revenue, brand strength | INMA

In a rapidly shifting global landscape, news media companies must rethink everything — and recognise that AI, diversified income, and brand strength will define the next decade of journalism. That was the opening message delivered by INMA Executive Director and CEO Earl Wilkinson at the Africa Media Revenue Summit earlier this week as he shared how the global media industry is undergoing a profound transformation — and offered Africa’s publishers best practices on how to navigate it.

Read more ▸

TV set is most popular way to watch YouTube in UK, study finds | The Guardian

The television has replaced laptops, tablets and smartphones as the most common device for UK viewers to watch YouTube at home, according to data confirming the platform’s place as a living room mainstay. More than half of all YouTube viewing through a domestic wifi connection is now done through the traditional TV, making it the top-ranking YouTube device across all age groups. The findings, from a Barb Audiences review, found that YouTube viewing is still skewed towards children, with whom the platform has been popular for some time.

Read more ▸

Vendor Integration Gives Cloud A Boost, But Broadcast Adoption Still Measured | TV News Check

Local stations turn to the cloud to launch FAST channels in search of new digital revenues, but have been slower to adopt it for traditional workflows. Over the past few months, a series of high-profile sporting events on OTT platforms, including a Christmas Day NFL game on Netflix, an NFL wild-card playoff game on Amazon Prime Video and both the Super Bowl and Winter Olympics on Peacock, have demonstrated the power of the public cloud to deliver live television programming at scale. But for day-to-day operations across broadcast networks and local stations, the adoption of cloud technology has been somewhat less dynamic.

Read more ▸

35-Year SEO Veteran: Great SEO Is Good GEO — But Not Everyone’s Been Doing Great SEO | SEJ

As SEOs, we are used to being adaptable to changing algorithms, so LLM optimization should be a simple extension of that process. To discuss the industry debates surrounding the differences between SEO and GEO and clarify whether they are the same or different, I spoke with SEO veteran Grant Simmons. Grant has over 30 years of experience helping brands grow and has spent decades focused on meaning, intent, and topical authority long before LLMs entered the conversation..

Read more ▸

A New Startup Is Personalizing Paywalls for Bots and Humans Alike | AdWeek

Stop me if you have heard this before, but it bears repeating: When OpenAI debuted ChatGPT in November 2022, it introduced an existential problem that, even after more than three years, has yet to find a solution. In short, answer engines rely on the copious availability of quality data to fuel their responses, which they draw from content creators, like publishers. Yet in answering users’ questions without sending those users back to the websites that supplied the data, they starve these publishers of revenue, creating a system that is, quite obviously, unsustainable.

Read more ▸

The authority era: How AI is reshaping what ranks in search | Search Engine Land

In the early days of SEO, authority was a crude concept. In the early 2000s, ranking well often came down to how effectively you could game PageRank. Buy enough links, repeat the right keywords, and visibility followed. It was mechanical, transactional, and remarkably easy to manipulate. Two decades later, that version of search is largely extinct. Algorithms have matured. So has Google’s understanding of brands, people, and real-world reputation. In a landscape increasingly shaped by AI-powered discovery, authority is no longer a secondary ranking factor – it’s the foundational principle.

Read more ▸