FIPP's latest data shows a digital subscription market growing 24.7% and concentrating fast. The winners are bundling, embedding, and reducing their dependence on search. The losers are still renting their audiences from Google.

Nine international publishers, Le Monde, El País, The Irish Times, Politiken, Corriere della Sera and De Standaard among them, now sell subscription bundles that include access to The New York Times alongside their own journalism, according to Nieman Lab. The arrangement suits both sides. Local publishers add internationally recognised reporting to their value proposition without the cost of producing it. The New York Times Company extends its subscriber reach into markets where direct acquisition is slower and more expensive.

One of those nine publishers, De Standaard, sit inside Mediahuis, whose “essential subscription” is the clearest worked example of the strategy. As I wrote for Audiencers, Mediahuis has built a bundle that wraps journalism, e-books, masterclasses, a hiking and cycling app, and access to The New York Times into a single subscription. The early data is compelling: activated users churn 26.5% less, and the bundle is engineered to push subscribers up from the basic tier to a premium one priced at roughly double. The New York Times partnership is one component of that bundle, not a curiosity bolted on top.

That arrangement is the clearest signal of where the market is heading. FIPP’s latest Subscription Report counts 53 million digital-only subscriptions across 206 titles, and the headline 24.7% growth conceals a sharp split. Publishers that have built direct audience relationships, bundled product ecosystems and cross-title access are accelerating. Those still dependent on single-title subscriptions and search-driven acquisition are losing ground, in some cases losing subscribers outright.

The rapid expansion of AI into search has sharpened the divide. Referral traffic that publishers relied on for audience acquisition and advertising revenue is compressing. Publishers that were renting their audiences from Google are finding the lease harder to renew. The response from the strongest performers is consistent: own the relationship, deepen the product, reduce dependence on external distribution.

No company illustrates that logic more clearly than the New York Times Company, which is why the cross-publisher bundles matter: they are central to the strategy, not a side experiment.

The New York Times is no longer just competing. It’s embedding

The New York Times Company now reports 12.21 million digital-only subscribers and stopped reporting title-level numbers this year. The second decision reframes what kind of business the company believes it is in. The New York Times is no longer a newspaper with digital products attached. It is a subscription business that happens to include a newspaper, alongside Games, Cooking, Wirecutter and The Athletic, bundled into a single proposition designed to make cancellation feel like losing several things at once.

The cross-publisher bundles extend the same logic outward. As AI search compresses referral traffic, the value of a subscription with multiple compelling reasons to stay increases. Bundles reduce churn inside a single portfolio. Cross-publisher bundles spread that logic across the whole market and position the New York Times Company as the connective tissue of other publishers’ subscription strategies, not just its own.

The Nordics settled this argument years ago

While publishers in the US and UK are still debating the merits of bundling, Norway’s Amedia has been running the experiment since 2020, and the results are stark.

Its +Alt bundle, which gives subscribers access to all 127 of Amedia’s local newspapers plus a sports streaming service, now has 433,000 digital-only subscribers. Churn on the bundle runs at 0.7%. Churn for a single-title subscription is 16.4%. That difference translates into a subscriber lifetime value that is 26 times higher for the bundled product.

Norway, as a subscription market, grew 55.56% in the period, a function of a deliberate structural choice. Publishers there decided early that local relevance and shared digital infrastructure were not in tension; they were complementary.

In Sweden, Bonnier News reached the same conclusion. Its +Allt bundle, combining national titles (Dagens Nyheter, Expressen), more than 40 local newspapers and magazine content into a single subscription, reached nearly 500,000 digital subscribers by the end of 2025, up from around 270,000 in mid-2024. Including print subscribers with bundled access, total reach is over 1.1 million.

Germany and France are also moving. Germany’s subscription market grew 35.83% in the period, the fastest of any major European market. France grew 31.21%. Both reflect the same underlying pattern: publishers deepening their product propositions and reducing their exposure to external distribution.

The UK is following. FIPP’s data shows the market up 22.61%, with Newsquest, DC Thomson and Mail+ among the standout performers on percentage growth. Mail+ grew 177.8% in the period, from a low base, but the direction is clear.

AI search didn’t break the business model. It revealed which models were already broken

The useful question is not whether AI search hurts publishers. It is which publishers. The FIPP data, read against recent industry reporting, makes the divide specific.

The content most exposed to AI overviews is the content most easily summarised in a sentence: car specifications, factual service journalism, commodity news. Bauer Media Group has said as much directly. When the overview answers the question, the click never happens, and the publisher’s role in the value chain ends with the model’s training set.

The content least exposed is reporting that does not compress. Analysis, investigation, distinctive voice, and depth all convey information that survives summarisation, because the value lies in the framing rather than the facts. The Telegraph’s SEO director has described search as being in “managed decline” rather than collapse, on the grounds that subscribers return for reporting summaries that cannot be replicated elsewhere. The framing fits the data.

The more interesting signal sits in the other direction. The Washington Post reports that readers arriving via large language models spend longer on the site and convert to subscriptions at higher rates than traditional search visitors. The volumes are small, but conversion quality matters: AI search may be filtering out low-intent traffic that inflated publisher metrics without ever converting into subscribers.

As Colin Nagy argued recently in Monocle, AI hasn’t created a new problem for publishing; it has clarified an old one. Publishers built around search traffic were never really building audiences; they were borrowing them, and the loan is now being called in. The publishers losing ground are not victims of a new technology. They are tenants whose landlord has changed the terms.

US local publishers are paying it

Lee Enterprises reported that its digital subscriber base fell from 728,000 in March 2025 to 609,000 by December: a loss of 119,000 in nine months. The USA Today Co, formerly Gannett, saw a 26% decline year-on-year, dropping from 1.95 million subscribers to 1.45 million.

Both companies attribute the declines to deliberate strategies focused on price optimisation and improving average revenue per user. That may be true. It also describes companies managing contraction rather than building growth.

The sharpest contrast sits within the same corporate structure. Newsquest, the USA Today Co’s UK division, grew digital subscribers 35% year-on-year to 139,000 over the same period. Same parent company, opposite trajectory.

What the winners have in common

Bloomberg grew subscription revenue 10% in 2025, taking paying subscribers beyond 707,000. It now produces 800 hours of video each month. Subscriber-only newsletters grew 20% year-on-year. Podcast downloads rose 26%. The pattern is consistent: more reasons to open the product, more often.

Substack, with five million paying subscribers, has built its growth engine around internal recommendations that move readers between newsletters inside the platform. It is less a publishing platform than a subscription marketplace, one that reduces its dependence on external traffic by creating its own internal referral economy.

Immediate Media, with 767,000 digital app subscriptions now accounting for 64% of group subscriptions, is using AI to identify which audiences are most likely to subscribe and when engagement signals are strongest. The objective is habitual use and longer subscriber lifetime value.

The thread connecting all of them is the same one running through the Nordic data, the NYT bundle strategy, and the divergence between US and European local publishers. Subscriptions built on borrowed traffic are fragile in ways that subscriptions built on daily habit, multiple products and direct relationships are not.

Growth in the market is concentrating around publishers who solved for retention before they needed to.

The post The New York Times is becoming subscription infrastructure for other publishers appeared first on Audiencers.

Hi, I'm Lennart Schneider, Founder of Subscribe Now, helping decision-makers in the subscription economy attract subscribers and keep them happy.

When I talk to newsletter creators, it usually doesn't take long before one of two questions comes up:

1) What actually constitutes a good open rate?

2) My newsletter has an open rate of 40%. That's great, isn't it?

The honest answer is: There is no such thing as a "good" open rate, and the open rate really says nothing about the quality of your newsletter. 

In this article, I want to clear up this misconception and explain when you should still pay attention to it.

P.S. I've written a book! The playbook for successful subscription models: With growth strategies and best practices from 50 leading companies. You can pre-order here (or on Amazon).

Why open rate says nothing about the quality of your newsletter

I completely understand the need for benchmarking and comparability. With newsletters, you have relatively few metrics, so you try to draw as many conclusions as possible from what you do have. Often, however, you draw more conclusions than are actually useful.

Let’s start with the basics: open rate is a fraction. You calculate it by dividing the number of unique opens by the total number of recipients. So far, so simple.

However, this leads to a small detail that is often overlooked. You can increase the open rate in two ways: more openers, or fewer recipients.

The latter point is often overlooked, but it’s crucial. People who don’t open your newsletter are often the same ones. And especially with older mailing lists, you often have a large proportion of inactive subscribers who have lost all interest.

So, if you have a 50% open rate with 10,000 recipients, then there are probably several thousand among them who haven’t opened a single email in six months. Let’s say there are 3,000 of them. If you remove them from the mailing list, your open rate immediately jumps from 50% to 71.4% (5,000 opens/7,000 active recipients).

Has this improved the quality of your newsletter? Are your readers more satisfied? No!

Has the quality of your distribution system improved as a result? Absolutely!

Why would you delete valuable leads?

Admittedly, I’ve recommended this to many newsletter creators, but the enthusiasm is usually lukewarm. They’ve put a lot of money and effort into acquiring these subscribers, so why would they just delete the addresses? Perhaps the reach is even important for advertising sales (even if nobody sees the ads).

But here are a few reasons why it’s worth it:

1. If someone hasn’t opened your emails for a long time, the likelihood of them returning is relatively low
2. You increase the risk of being perceived as spam
3. Depending on the newsletter tool and contract, costs increase as you send more emails
4. Your brand suffers when customers perceive you as pushy
5. Your advertisers are surprised when clicks fail to materialize on a supposedly large distribution list, and become suspicious
6. …

Many of the world’s leading newsletters therefore place great importance on the quality of their mailing lists and remove users after a long period of inactivity. The New York Times even discloses this quite transparently:

How can I find out who is inactive?

Good question! And not so easy to answer. Since Apple’s Mail Privacy Protection, it’s often difficult to track who actually opens an email – Apple automatically opens emails before they reach the recipient, and therefore your email program thinks these users are active. Other users block tracking and are registered as inactive, even though they enthusiastically read every issue. So if you’re unsure: just ask the users.

Clicks can be tracked more reliably than opens, and if readers don’t click after (repeated) requests, you can remove them from the mailing list with a clear conscience.

Do I really need to delete them?

No. You can also pause them for now, or reduce the frequency. A good example is the sports newsletter “The Gist”. They temporarily pause inactive subscribers, and then when a major event is coming up (for example, the Olympics), they try to reactivate them.

However, you shouldn’t overdo it. If you haven’t contacted them for 1.5 years, you should continue to refrain from doing so. After that time, their consent to contact them also expires.

Even more reasons why open rate is misunderstood

Your open rate depends on numerous factors, and the quality of your content is just one of them. Here’s a (likely incomplete) list of factors that affect open rates:

Mailbox display:

• Subject lines (Do they encourage clicks?)
• Pre-header (Do you tease the content well?)
• Sender’s name (Do they trust you and look forward to your emails?)
• Sender images (Will they stand out in the inbox? Only works with certain clients)

The quality of your distribution:

• Cleanup of inactive users (Are non-openers regularly removed?)
• Segmentation (Do you always send emails to the entire mailing list or do you select recipients based on interests?)
• Preference Center (Can users configure which emails they receive and how frequently?)
• Age of the addresses (How long ago did someone sign up?)
• Lead campaigns/address origin (On which channels and with which promises were the users acquired? Did they want the newsletter, or did they just share their email to participate in a competition, for example?)

Deliverability

• Advertising tab in Gmail (Are emails from Gmail classified as advertising and not delivered to the main mailbox?)
• Spam (Do the emails often end up in the spam folder?)
• Bounces (Can the emails not be delivered? What are the different types of bounces, hard and soft?)
• Image sizes (are the images too large?)
• Email size (Will the email be truncated, for example, by Gmail, which happens from 102 kb upwards?)
• Shipping time (Is the shipping time chosen?)

Contents

• Good quality & relevance
• Diversity (do you cover different needs so that every issue contains an “aha” moment?)
• Continuity & predictability (Do readers know what to expect and why each issue is worthwhile?)

Technology

• Email tool or measurement technology (each mailing tool measures opens slightly differently, which affects the KPIs)
• Tracking Opt Out (Have users disabled tracking of clicks and opens?)
• Auto Opens on iOS (Are open rates inflated because emails are automatically opened due to privacy settings?)

So I can ignore the open rate?

Not quite. Since we don’t have many metrics, it’s still a valuable signal, as long as you’re aware of what it tells you, and what it doesn’t.

What you shouldn’t do is compare your absolute open rate with competitors who use completely different tools and whose lead generation works differently. This apples-to-oranges comparison is completely pointless.

What you can do instead:

• Observe long-term trends: Is my own open rate constant? Have there been any sudden changes I should understand? (But keep in mind that these changes could be due to technical reasons, e.g., more users with tracking protection.)
• Optimize your open rate with A/B testing: e.g., send each email with 3 different subject lines to a sample group beforehand and send the best version to the rest of the distribution list.
• Examine outliers: Were there any particular issues that were opened more or less frequently than average? What can I learn from them?
• Re-contacting those who didn’t open the email: Some companies send a newsletter a second time if it wasn’t opened the first time. I find it a bit spammy, but it seems to work.
• Use net reach (recipients x open rate) as the basis for advertising deals: This is fair to your advertising clients and you can also justify why the CPM (cost per thousand contacts) is higher if you are contacting a cleaned distribution list.
• Comparing segments: When running different lead generation campaigns, you shouldn’t just focus on the short-term cost per lead (CPL), but also consider whether the campaigns generate active recipients in the long run. A sweepstakes often generates a large number of addresses cheaply, but the open rate drops significantly after just a few campaigns.
  For better comparability, you can calculate a cost per lead after, for example, 10 campaigns. If you generate 1,000 leads per campaign for €1,000 in each of two campaigns, the short-term CPL is €1. If, after 10 campaigns, 70% of the leads in campaign A are still active, while only 30% remain active in campaign B, you get a clear picture: In campaign A, an active lead cost €1.43, while in campaign B it cost €3.33.
• …

PS: My newsletter has an open rate of 51-58%. That’s great, isn’t it?…

PPS: To hear more from me about subscription and newsletters, sign up to my own newsletter, Subscribe Now, here

This article was originally published in German on the Subscribe Now website, translated and republished with permission.

The post Why I don’t care about your newsletter open rates appeared first on Audiencers.

Regula Marti helps media organizations navigate transformation at the intersection of editorial, product, technology, and business strategy.

In this article, Regula discusses how engagement isn’t a single metric to maximize — it reflects different reader needs and moments, where the same behavior can signal success or friction depending on context. She recommends moving beyond averages and article-level metrics to analyze engagement by audience segments and sessions, focusing on whether readers get what they came for and return.

Most publishers have more engagement data than ever — and still struggle to agree on what engagement actually means. The tools aren’t the problem. The problem is the assumption behind them.

The assumption is this: engagement is one thing, and more of it is better. But engagement isn’t one thing. It’s the outcome of different reader behaviors, driven by different motivations.

Same reader, different moments

Think about how you consume news. Some mornings you want a quick overview. Some evenings you want to go deep. Sometimes you’re browsing. Sometimes you want distraction.

These aren’t different readers. They’re the same reader in different situations, with different needs.

A short session can mean the reader got exactly what they needed. A long one can mean depth — or friction. High article counts can mean curiosity, or confusion. The same metric means something different depending on what the reader was trying to do. Without that context, you’re pattern-matching on behavior you don’t fully understand.

Who you’re actually looking at

Here’s a pattern that shows up consistently across publisher data: a small group of heavy users (20+ visits per month) drives a disproportionate share of pageviews. A large flyby majority (one visit per month or fewer) makes up most of the audience but a tiny fraction of consumption. Frontpage clicks illustrate this well: they overrepresent heavy users. Optimizing for clicks risks building a homepage for the audience you already have, not the one you’re trying to grow.

In between sit the loyal readers, visiting 2 to 20 times a month. They have a relationship with you but haven’t yet built a habit. They’re the most likely to convert and retain — and the easiest to overlook, because heavy users dominate the averages. The real movement happens here: turning flybys into returning readers, and loyals into heavys.

What a summary feature taught us

Product decisions show what happens when you measure at the wrong level.

When we introduced article summaries across our portfolio at Tamedia — short digests at the top to help readers decide whether to go deeper — the concern was obvious: readers would skim and leave. That’s partly what happened. Scroll depth dropped for some users.

But at session level, something else emerged. Those readers went on to read more articles per visit, and overall session time increased. The summary helped them get oriented quickly and navigate to what mattered for them more in this moment.

Other readers used it differently. After reading the summary, they were more likely to finish the full article. It gave them confidence it was worth their time.

Two opposite behaviors at article level. Both positive at session level. Both pointing to the same outcome: a need met, and a higher likelihood to return.

Segment awareness can also shape product design from the start. We kept our fast news format intentionally compact. Heavy users quickly find what they need, while the freed-up space showcases the differentiated content that loyals and flybys need to keep coming back.

What this means for how you measure

The answer isn’t a more sophisticated metric. You need to be more deliberate about what you measure and why.

Look at your segments separately in your analytics tool. The same metric tells a different story across them — and the average hides most of it. Track them over time: a single snapshot tells you little; the direction of travel tells you a lot.

Article-level metrics will often mislead you. A reader can leave early and still have had a successful session. That’s why session-level metrics matter — but only when read together with segment data. A daily digest designed for quick orientation will show low time-on-page. That’s not underperformance — it’s the format working as intended. The right metric is whether those users come back the next day.

Try this in your next team meeting

Start in your next team meeting. Pick one feature or format you’ve recently launched or are evaluating. Review its impact at article level, then at session level. Then ask: for which segment was this actually useful — and in which moment? A multi-format feature is a good example: a reader who skipped the article but listened to the audio version on their commute — possibly in a podcast app, invisible to your analytics — may well return the next morning. Low article engagement, high likelihood to come back.

You’ll often find that what looks like low engagement in one view is exactly what success looks like in another.

The reader who got what they came for is more likely to return. Design your metrics to capture that.

The post The reader who left your article early might be your most engaged user appeared first on Audiencers.

Max Moné is co-founder and CEO at Poool, the dynamic journey builder to boost subscription conversion, engagement, and loyalty.

This is the fifth in a 6-part series where I share what I learned from studying 100 subscription business models across 15+ industries. 
> Episode one: 100 subscription business, 15 industries, 1 moodboard
> Episode two: What’s free, what’s paid, and why it’s really an engagement trade-offf
> Episode three: Why the best subscription businesses don’t try to sell on day 1
> Episode four: The subscription pricing page we've all built at least once

The cycle doesn’t stop when someone becomes a subscriber

In episode 1, we introduced the idea that the best subscription businesses don’t follow a linear journey (attract → convert → retain) but an infinite cycle of micro-conversions, where each step has one goal and unlocks the next. In episodes 3 and 4, we looked at what happens before the first paid conversion: how these businesses use registration, onboarding, and pricing to guide the user step by step.

This episode covers what happens after someone becomes a subscriber.

And what we observed is that the journey really doesn’t change in nature once the user has paid. After someone subscribes, the next micro-conversion might be signing up for a newsletter. Then downloading the app. Then discovering podcasts or games or a specific section of the product. Then, when the subscriber is ready (and when you know enough about their usage), you can propose an upsell or bundle.

Acquiring a subscriber isn’t the end of the funnel. It’s just one more step in the infinite cycle we described in episode 1. And what most people call “retention” is really just the continuation of the same sequence of micro-conversions, except now you have data about the user and can be much smarter about what you propose next.

The problem is that most of this work needs to happen fast.

Why you have about 100 days (and probably less)

INMA data shows that a quarter of subscribers cancel in the first month, and half are gone by the fourth month. And a Nieman Lab study on The Wall Street Journal’s data showed that the probability of adopting a new habit drops dramatically after about 100 days.

Half of your new subscribers are gone within 4 months, and after about 100 days the window to build a new habit narrows significantly. If you don’t manage to execute enough micro-conversions during those first few months (newsletter signup, app download, feature discovery, habit formation), the probability of keeping that subscriber drops, regardless of how good your anti-churn campaigns are later on.

(I know this isn’t the most cheerful insight we’ve had in this series, but it’s what the data consistently shows.)

What this looks like in practice: L’Équipe after the Olympics

This is where the theory gets concrete. L’Équipe (a Poool client), speaking at The Audiencers’ Festival in 2025, shared how they built 12 post-subscription projects after the Paris 2024 Olympics, which had generated a huge spike in new subscribers, subscribers they needed to keep. I won’t go through all 12 (the full article is worth reading), but the ones that stood out to me are the following.

Their data analysis revealed something I found really striking: what a subscriber does after paying has more impact on M+1 churn than what they did before subscribing. When Fabien Mulot showed this at The Audiencers’ Festival, it was one of those moments where you think “of course, that makes total sense” and at the same time you realize that most of the industry’s energy is still focused on acquiring subscribers, not on what happens once they’ve subscribed.

Based on this insight, they focused their onboarding on one specific goal: getting new subscribers to read at least two paid articles within the first three days. Their data showed that each paid article read during those first 72 hours reduces the following month’s churn by about 5 points, so two articles in three days already makes a meaningful difference on the curve.

The onboarding sequence they built follows the same micro-conversion logic we’ve been describing throughout this series:

• first, welcome the subscriber and showcase what their subscription unlocks;
• then encourage them to personalize their experience (alerts, newsletters, favorite topics);
• then immediately serve them paid content tailored to their interests. Each step has one job.

The cumulative impact is significant.

• Simply being exposed to this onboarding journey reduces churn by 7 points at M+1.
• Completing the onboarding journey while also activating an action (turning on alerts, signing up for a newsletter) reduces churn by 20 points at M+3
• The more micro-conversions a subscriber completes in those early weeks, the less likely they are to leave.

Beyond onboarding, L’Équipe continued the logic with a newsletter strategy (a monthly preview of upcoming content, plus journalist-led newsletters across four sports verticals where subscribers can ask questions directly to editorial teams), and a new “Games” vertical to expand consumption into other formats. Each of these is another micro-conversion in the sequence, another reason for the subscriber to come back tomorrow.

El País and FT Strategies confirm the pattern at a broader scale

L’Équipe’s data is specific to sports media, but the pattern holds across other publishers.

El País shared at a 2024 INMA conference that subscribers who use editorial newsletters churn 24% less, app users churn 15% less, premium newsletter users churn 29% less, and loyalty scheme users churn 30% less. Each one of these is a micro-conversion that happened after subscribing, and each one reduces churn by 15 to 30%.

FT Strategies ran some very complementary work, analyzing the impact of each product feature on engagement vs. the percentage of subscribers who actually use it. The features with the highest impact (multimedia, interactive content, games) are also the ones that fewer than 30% of subscribers use. The gap between impact and adoption is where the biggest opportunity sits.

What happens when you get the micro-conversions right: lock-in effects

When the micro-conversion sequence works well over time, you eventually reach a point where leaving becomes genuinely hard for the subscriber. This is thanks to long-term engagement and relationship-building.

But of course, there are other strategies to keep your subscribers retained, ones that can be used in conjunction with the micro-conversions detailed above. When combined, they’re powerful for developing loyalty.

During our benchmark, we saw three main types of lock-in across the 100 businesses:

• Personalization – we talked about Spotify in episode 4, but the principle applies to any product that gets smarter with use, adapting the product to the user
• Multi-accounts – Netflix profiles, Blinkist’s partner account where you sometimes keep the service for the other person
• Bundling with external services – Canal+ with Netflix, Amazon Prime combining shopping, streaming and delivery, Walmart+ bundling grocery delivery with Paramount+

(You can read more about lock-in effects here)

Funnily enough, each of these is really the result of many successful micro-conversions accumulated over months and years. The lock-in is earned, not imposed.

Anti-churn: necessary, but it’s the safety net

And then there’s anti-churn, which is what happens when the micro-conversion sequence hasn’t worked well enough and the subscriber is about to leave.

L’Équipe has been methodical about this too. Their cancellation form (which handles two-thirds of all unsubscriptions) was redesigned through A/B tests: they added an intermediate screen summarizing the benefits the subscriber would lose (+2% retention), and tested personalizing the message with the subscriber’s first name and a specific discount percentage (+1.2 points on conversion). Small changes, but measurable impact at scale (source).

Unidad Editorial took a different approach by building a churn propensity model that predicts which subscribers are about to cancel, triggering personalized actions before they reach the cancellation page. Result: a 25% reduction in churn in three months.

These are well-executed tactics. But in the broader picture, they’re the safety net.

The real strategy is everything that happens before the subscriber even thinks about leaving: the onboarding, the feature discovery, the newsletter signup, the app download, the habit formation. All of those micro-conversions that, together, make the product too valuable to walk away from.

What you can take away from this (and test tomorrow)

Map the post-subscription micro-conversions you want your subscribers to make. Newsletter signup, app download, podcast discovery, games, alerts configuration, etc. based on their value. Then ask yourself: do you have a sequence designed for this, or are you hoping it happens on its own?

Focus on the first 3 days and the first 30 days. L’Équipe’s data shows that each paid article read in the first 72 hours reduces the following month’s churn by 5 points, and a completed onboarding with an action reduces churn by 20 points at M+3.

Check the gap between feature impact and feature adoption. FT Strategies’ framework is a useful lens: if your most impactful features are also the least used, that’s your biggest opportunity. And El País data shows that each additional touchpoint is worth 15 to 30% less churn.

Treat anti-churn as a safety net, not as a strategy. If the majority of your effort goes into save campaigns for subscribers who are already about to leave, you’re fighting the battle at the wrong end of the timeline.

And if all of this sounds like a lot to orchestrate (because it is): that’s exactly the kind of problem we’re solving with Poool and Darwin CX. From onboarding sequences to engagement actions to anti-churn triggers, across web, app, and print. Orchestrating micro-conversions at every step of the subscriber journey is what we do. If that resonates, or if you’re currently looking at your subscription platform options, let’s talk.

Next week, episode 6, the last one: toilet paper, Japanese snack boxes, razor blades, wine, and genealogy. The most surprising subscription models we came across, and what they confirmed about engagement (because at this point in the series, I think you already know what the answer is going to be).

The post The first 100 days: why most of what we call retention is really just micro-conversions (again): episode 5 appeared first on Audiencers.

Audience segmentation is more than a technical exercise, it is a critical strategic driver for any modern media organization. In a landscape where generalist appeal is fading, dividing your audience into distinct groups based on shared characteristics allows you to move beyond guessing and start delivering real value.

At Audiencers’ Festival in Hamburg on March 3rd 2026, subscription consultant Selma Stern shared her recommendations for mastering audience segmentation to maximize the value of each reader.

Why prioritize segmentation?

Segmentation can be mind-bogglingly complex and wildly impactful.

The word comes from biology, where it refers to the process of cell division from a fertilized egg to a full organism. In business, it refers to identifying the constituent parts of a whole – ideally, a set of mutually exclusive and collectively exhaustive groups of customers. It can’t be perfect by definition, but even an imperfect segmentation can massively improve your bottom line.

At its core, segmentation addresses three fundamental business needs:

• Stickier content: By clearly defining  Ideal Customer Profiles (ICPs), newsrooms learn not just what to write about, but how to speak to their readers so the content truly resonates. If you replicate your ICPs as agents and train your newsroom to use them, you can get endless customer feedback at any time of the day.
• Smarter monetisation: Without segmentation, a dynamic paywall is only based on recency and frequency – which is fine, but does not capture your full market potential. . With it, you are “fishing where the fish are,” targeting the right users at the right time.
• Better marketing: Different segments live on different channels and respond to unique triggers. Segmentation tells you exactly where to find your users and helps advertising brands connect with theirs.

Choose your lens:

To build a complete picture of a user, like “Amanda, 40, VP of Marketing,” organizations can look through several analytical lenses:

Segmentation type	Focus/use case	Signal examples
Behavioral	UX Optimization	>5 visits per month, scroll depth, newsletter opens
Needs-based	Product Development	Prefers summaries, explainers, or opinion pieces
Value-based	Revenue Optimization	LTV, subscription tier, payment history
Technographic	Paywall Optimization	Device type (Mobile vs. Desktop), OS, Referral (LinkedIn vs. Facebook)
Demographic	Ad Sales / Strategy	Age cohort, geography, seniority

> Spektrum’s evolution of audience segmentation and testing

Five tips for succeeding in segmentation

1 – Hire strong analytics: This is the “single most important lever.” You need talent that can bridge the gap between high-level business strategy and deep data modeling.

2 – Make it cross-functional: Segmentation is a “Left Brain + Right Brain” exercise. Editorial, product, marketing, and data teams must own the model together; no single department can succeed in a silo.

3 – Quantify benefits early: You must build a business case or risk losing organizational priority. Use “Napkin Math” to show the potential uplift: for example, if a newsletter with a defined audience grows twice as fast as a vague one, what could the vague one be worth with a properly defined audience?

4 – Remember the customer: Segments should feel like people, not rows in a spreadsheet. Give them names like “Jane, the CMO” or “Joe, the Sales Lead.” Talk to them, build agents around their profiles, and ask, “Would Jane actually read this?”.

5 – Keep it simple (but useful): Aim for 4–6 mutually exclusive, collectively exhaustive (MECE) segments. If your team can’t name the segments from memory, your model is too complex and will fail to be actionable. Even if your goal is to build a dynamic paywall algo – lose (human) control of complexity, and you won’t know what’s working, which means you won’t be able to adjust next time the world changes.

Real-world impact: fewer paywalls, more subscription revenue

A US publisher’s journey from a rigid freemium model to smart segmentation highlights the power of behavioral data. They discovered that just 1% of traffic drove 25% of conversions.

By analyzing attributes, they found that newsletter click-throughs and direct repeat traffic were high-value, while Facebook and Android traffic converted at near-zero rates. By stopping the paywall for near-zero segments and doubling down on high-value behavioral and technographic signals, they achieved:

• A 2x increase in conversion rate.
• 50% fewer paywall exposures, leading to a better user experience for the same number of new subscribers.

Did this publisher get it completely, fully right? Absolutely not. But the model was a dramatic improvement over the simplistic frequency and recency model they were using before.

As statistician George Box famously said: “All models are wrong, but some are useful”. Be pragmatic—start with the data you have, build empathy for your readers, and iterate until your segments drive measurable growth.

The post Mastering audience segmentation with Selma Stern appeared first on Audiencers.

In this guide, we share the various ways that publishers can segment their audience to develop a dynamic paywall model, with examples for each type of segmentation: 
> By content type: free vs premium, user needs, date of publication...
> By user profile: level of engagement, user status, location...
> By acquisition channel: Google, social media, newsletters...

Some quick definitions: 
> Audience segmentation: grouping users based on their profile or context, i.e. data that could impact their behaviour, inform us on what could convince them to subscribe or make them more or less likely to convert in that moment
> Dynamic paywall: a paywall that adapts for different audience segments. This could be in terms of the paywall design, messaging or even the journey itself (e.g. seeing a paywall or not). Note that "dynamic" doesn't necessarily equate to AI-driven. It's often rule based - i.e. a user reading a "politics" article sees a different paywall to a user reading a "sports" article.

There are 3 main buckets of audience segmentation with the goal of increasing conversion rates:

• By content type
• By user profile
• By acquisition channel

In each of these cases, the form of segmentation dictates who sees or experiences what. For instance, if we segment based on location, someone in Germany might see and experience something different than someone in the UK.

What can change based on the segment, in terms of the conversion strategy?

• The fact that the article is blocked or not
• The conversion journey (e.g. how many articles for free before the paywall, the use of a registration wall)
• The wording on the paywall
• The design (colours, image, etc.)
• The pricing and subscription offers pushed

Let’s dive into these 3 buckets of segments, with examples for each:

1. Segmenting by content type

For many publishers, this is the simplest starting point. It allows for extensive testing to discover what content converts best without risking overall traffic. It also helps to put editorial teams in control (to an extent) as the paywall strategy is based around their work instead of a user’s profile. This can be valuable when first launching a dynamic paywall to get buy-in from these teams.

This form of segmentation can take various forms:

• Content divided into free (often general news or commodity pieces) and premium (generally longer-form, investigative articles which can’t be found elsewhere)
• Format-based: e.g. articles vs games vs videos
• Topic- or tag-based: sports, politics, news, lifestyle
• User needs based: update me, give me a perspective, etc.
• Date of publication: e.g. articles published more than 30 days ago

Free vs premium

Organising content based on its value:

• Free content: keep open to maximize reach (often focused on ad revenue)
• Premium content: close to convert readers into subscribers

Some publishers are developing this a step further by adding 2 additional segments:

• Grey content: sometimes open, sometimes closed (based on editorial decision)
• Super Premium Content: with a hard paywall

This segmentation is often driven by two business goals: advertising and subscriptions. However, the articles used for advertising revenue (which aim to gain page views) can also serve the role of engaging audiences to move them through the funnel toward subscription.

A good place to start is a matrix analysis, placing articles in different buckets based on their “strengths”. A goal can then be established for each bucket, making the most of each article.

For instance, Mather suggests the following matrix:

• Mass appeal / top of the funnel: advertising revenue focus, so left open
• Premium content / bottom of the funnel: subscription focus, so behind a paywall
• Mission journalism: often left open
• Under-performers (that require further analysis)

A slightly more complex approach is illustrated by FT Strategies below, who established 4 categories based on topic popularity vs willingness to pay:

The post Guide: how to segment audiences for a dynamic paywall appeared first on Audiencers.

Max Moné is co-founder and CEO at Poool, the dynamic journey builder to boost subscription conversion, engagement, and loyalty.

This article is the second in a 6-part series where Max shares what he learned from studying 100 subscription business models across 15+ industries. If you haven't read the first one yet, start here. 

Why we’re focusing on media for this one

In most subscription businesses, the rules for accessing the paid offer are simple. You subscribe to a box, you get the box. You hit a usage limit on Claude, you’re asked to upgrade. You want to get access to Calm for guided meditation? You get 1 week, then have to pay. Each model has its logic, but it’s tied directly to the product.

In media, it’s different.

Media online was historically free (almost always). Then publishers moved to paid models, and products got more complex: apps, newsletters, podcasts, games, bundling, community. Because of this, an enormous variety of access models emerged.

So we decided to go deep into media alone in today’s episode. Because the question “what’s free and what’s paid” is one of the most consequential product decisions in our industry.

And there’s no right or wrong answer. The right model depends on your organization, your editorial DNA, your content, your business model… It’s specific to each media, each newsroom, each audience.

The goal here isn’t to tell you what to do. It’s to show you the range of what exists, so you can figure out where the opportunities are for your unique context.

The engagement-frustration trade-off

The data on this has been clear for a while now.

In 2021, we published the Digital Media Review (DMR – a research initiative with Google and Le Geste – a french association for publishers) at Poool. One data point stood out: the relationship between the share of traffic exposed to paid content and conversion. The correlation is clear: the more you expose, the more you convert. But only up to a threshold. Past that, conversion plateaus or falls. Frustration alone is not a sufficient reason to subscribe.

And the other side of the coin is just as important: the DMR also showed a strong correlation between traffic on paid articles and the share of users who simply leave the site when hitting a paywall. The more you frustrate, the more people bounce. That’s not just a missed conversion. That’s a direct hit on engagement, on page views, ad revenue, and your ability to build any relationship with that user.

In the same year, Mather Economics published a study (part of a GNI LatAm Subscriptions Lab report) looking at paywall hits per user. Subscription probability follows a bell curve: up with each hit, peaks, then back down. Same conclusion, different data.

Then in 2022, the New York Times shared what brought all of this together (Audiencers analyzed it here). When they launched their paywall in 2011, the meter limit was the same for everyone: a fixed number of free articles, a regwall, some more articles, and finally a paywall.

Over the years, they developed the Dynamic Meter, a machine learning model that sets personalized meter limits per user.

The core insight: the model optimizes for two metrics simultaneously, engagement and conversion. And these two metrics have an inherent trade-off. More paywalls → more subscriptions, but less readership. They proved this with randomized control trials: as the meter limit goes up, engagement increases but conversion drops.

With a fixed approach, you’re forced to choose. Maximize engagement, or maximize conversion. Not both. There’s always a window of lost revenue that a single rule can’t capture. The Dynamic Meter expands that window by personalizing: more frustration for high-propensity users, more openness for low-propensity users.

None of this is new. DMR: 2021. Mather Economics: 2021. NYT: 2022. We’ve known this for years. And yet, the vast majority of publishers still use a uniform paywall for all users.

If you don’t personalize, you’re forced to make a choice between engagement and frustration. And that choice won’t be the right one for all users. That’s why more and more sophistication has developed. And the jump from “manual” to “1-2-1” isn’t binary. There’s a whole spectrum in between.

---

The 6 levels of paywall sophistication

Being at level 1 doesn’t mean you’re behind. But it means there are probably missed opportunities. The question is whether these missed opportunities are big enough to justify the resources needed to reach the next step.

• Level 1: Fully manual

Free and paid decided by hand, article by article. An Atlas & Audiencers’ study confirmed this is still the norm: the decision is based on the article’s perceived value. Breaking news free, analysis behind the wall (to simplify it to its core).

• Level 2: Manual + automated rules. Paid content still decided manually, but rules modify access based on additional criteria.

Foreign Policy shared their approach on The Audiencers. They segment by publish date (under 7 days: hard-walled; older: standard meter), but also by cohort. Users with high search intent and churned subscribers are hard-walled (high subscription propensity). Social referrals see a registration wall with a free page view (low subscription propensity, but conversion on registration is 100x higher than on subscription). That registration data became their strongest conversion driver. No AI needed. Just smart segmentation.

• Level 3: Manual with override.

The base decision is manual, but the system can override contextually: closing a free article for a high-propensity user, opening a premium one for a segment you want to engage. A first step toward personalization without losing editorial control.

• Level 4: Manual + a third automated category.

Free and premium decided manually, but a “grey zone” is managed automatically based on user profile, device, or source. Ideal for publishers in transition between fully manual and more automation, while still giving the editorial team a great deal of control (yes, the business side gets some autonomy, but only on categories validated with editorial).

• Level 5: Fully automated for conversion and engagement.

The system decides everything: whether to show a wall, which type, what meter limit, per user. The engagement/frustration trade-off managed dynamically.

Business Insider does this with AI: reading habits, traffic source, content genre propensity. The algorithm decides: paywall, registration wall, or nothing.

The NYT does it with the Dynamic Meter (as we talked about earlier). It learns from first-party engagement data (no demographic or psychographic features), adjusts per user, and continuously runs randomized trials to improve. This is the model that expanded the opportunity window we described above (source).

• Level 6: Fully automated for conversion, engagement, AND advertising.

The decision also integrates ad revenue. Fortune’s CCO at the time, Selma Stern, explained that if a user’s subscription probability is too low (first-time visit, smartphone, social media), the paywall opens and the page view is monetized through ads. The goal is to maximize total user value, wherever it comes from (which I personally think is an amazing model for short term profitability – because yes, sometimes long-term profitability might require less revenue in the short-term).

---

What you can take away from this (and test tomorrow)

Figure out which level you’re at. If all your users see the same paywall regardless of behavior, source, or profile: what are you missing?

Try one rule-based variation before thinking about AI. Foreign Policy’s 100x stat didn’t require machine learning. It required a hypothesis and a way to test it.

Start by looking at your bounce rate on paywalled content. If a significant share of users leave the site when they hit your paywall, that’s revenue you’re losing twice: no subscription AND no engagement. The DMR, Mather Economics, and NYT data all showed the same thing: a uniform paywall always forces a trade-off. Even a simple first step (like showing a registration wall instead of a paywall to low-propensity users) can start closing that gap. That’s what Poool makes it easy to test.

See you next week for Article 3, where we leave media and look at how the best subscription businesses across all industries think about conversion.

---

PS: This article was about the decision logic (who sees a paywall, and when). We didn’t cover the blocking method: how the paywall is technically implemented, front-end or server-side, and what that means for content protection, AI scraping, and SEO. That’s a separate topic, and we covered it here: “Paywalls & SEO“. All 6 levels above work with all blocking methods.

The post What’s free, what’s paid, and why it’s really an engagement trade-off: episode 2 appeared first on Audiencers.

According to the 2025 Reuters Institute Digital News Report, digital subscriptions have hit a plateau. Across 20 countries, only 18% of people pay for online news, a figure that has barely budged since last year. So if you feel like you’re hitting a ceiling with your core audience, you aren’t alone.

However, in their latest report on the power of registration, the 2025 FT Strategies and Google News Initiative (GNI) Subscriptions Academy found that the most successful publishers are stoping the “anonymous-to-subscriber” leap and focusing on the “middle ground”: Registration. By converting anonymous visitors into known users through a lightweight value exchange, you aren’t just collecting emails, you’re building the resilience needed for a privacy-first, platform-fragmented market.

The post Building resilience with registration appeared first on Audiencers.

The Next Gen News Report 2 (NGN2), a collaboration between FT Strategies, Medill Knight Lab, and the Google News Initiative, provides a roadmap for news organizations to reach younger audiences. The core of the report is the 7 Modes of Engagement, which distinguishes how "next gen" consumers (ages 18-25+) discover, consume, and share news.

To reach the next generation of news consumers, publishers must stop thinking about “the audience” as a monolith and start designing for specific “modes” of behavior. The second Next Gen News report reveals that younger consumers are not passive; they are active “information curators” who cycle through seven distinct mindsets.

Let’s breakdown the 7 Modes of Engagement and uncover how you can win in them.

1. The “Sift” Modes: how news is discovered

In sift modes, users filter through their chosen information environments to separate news from noise. 

Across all markets studied, the report finds that young consumers actively curate their information ecosystems, seeking to reduce the daily mental load from sifting through endless content. Consumers intentionally mold their environment to refine and filter the news sources they encounter.

Observing this shaping process in the field showcases how sifting is richer than originally thought, now defined through three distinct modes of news discovery: 

Mode 1: Scroll

 The vibe: passively encountering news whilst browsing social feeds
 The energy: engagement is often spontaneous and fleeting, shaped by the emerging news producers they follow and platform algorithms.
 The strategy: intentionally create content that feels native to the platform and find patterns that break through to capture fleeting attention in the few seconds before viewers scroll to the next post

How to win at Scroll mode:

• Post consistently and predictably 
• Balance the tone to make news entertaining without trivializing serious events
• Develop repeatable storytelling templates.

Case Study: Local News International. Using “Dave,” a recurring character who plays the role of the viewer, they use skits and satire to cover complex news in 40 seconds. By using a consistent personality and humor, they stop the scroll and build a recognizable brand identity without formal news constraints.

Mode 2: Seek

 The vibe: actively looking for information on a specific event or topic
 The energy: a highly intentional mode of engagement, where consumers look for efficient, information-rich content that delivers depth and relevance without distraction
 The strategy: anticipate information needs and make depth accessible without friction, ensuring that credibility, clarity and efficiency guide every interaction.

How to win at Seek mode

• Guide discovery
• Deliver curated briefings
• Offer live formats that capitalize on immediate interest.
• Create seamless, intelligent discovery experiences (such as Ask FT or Ask The Post)
• Design intuitive interfaces within owned and operated platforms
• Provide high utility & control, such as personalization and tailored formats based on preferences, such as listening or watching

Case study: Particle News. Using AI, Particle gives readers the option of discplaying a news story in 6 different ways, including the 5 Ws, opposite takes on the same story or “Explain like I’m 5”

Mode 3: Subscribe

 The vibe: receiving news from trusted, curated sources (newsletters, podcasts, SMS) they have deliberately chosen
 The energy: people expect timely, important or novel updates from sources whose judgement they rely on and do not want to miss
 The strategy: timely, important or novel updates from reliable sources

How to win at Subscribe

• Sync with daily habits: design delivery around audience routines and preferences, not output cycle, and consider offering personalized delivery
• Make subscribe content dense and easy to complete
• Standardize the skeleton for instant scanning
• Keep it short, but provide options to go deeper

Case Study: The Edinburgh Minute. Short paragraphs, plain language, softer tone in the introduction. A long list of key stories, each with two sentences max, along with an emoji to introduce the topic, and a follow-up link.

---

2. The “Consumption” Modes: how news is processed

In consumption modes, users have chosen to explore a subject, topic or piece of content more deeply.

Audiences are more likely to engage further when a story feels personally relevant, emotionally resonant or sparks their curiosity. How they first encounter that story does not dictate what happens next. Differing motivations shape how individuals choose to go deeper, whether by verifying facts, exploring context, or interpreting different perspectives.

The 2025 findings reinforce the three modes we previously identified that capture how audiences engage more deeply with the news once they move beyond sifting. These modes are:

Mode 4: Substantiate

 The vibe: verifying if a story is true or finding the "raw" facts.
 The energy: an individual has found a piece of information and wants to quickly and credibly confirm it
 The strategy: create content that is promptly delivered after a breaking news event, focusing on main claims and facts, with simple, clear and digestible evidence at its core

How to win at Substantiate mode:

• Meet your users where they are
• Prioritize accuracy and relevance over speed or volume
• Push information directly to audiences
• Lead with facts
• Create quick, easily digestible formats 
• Show your working: stitch primary source materials into the content and show the process

Case Study: UnderTheDeskNews. They often presents sources behind the presenter, such as maps of military facilities, to promote substantiation. These successful news producers clearly annotate from where information has come using overlays on videos or references in text, and they include links in captions and bios

Mode 5: Study

 The vibe: going deep to master a topic of personal or professional interest
 The energy: a desire to learn, upskill and be inspired. For some, Study mode is a pleasant distraction
 The strategy: producers can create long-form, information-dense content that has high-production value to keep engagement high, while showcasing their lived experience as a means to draw users into the story.

How to win at Study mode:

• Bring the story to life through immersion.
• Show interest
• Open with the motive, not just the topic
• Use mini-brands or shows as a creative sandbox
• Start with a learning goal to orient the audience and prime attention
• Favor evergreen or under-explored themes
• Build concepts progressively and go deep
• Bring it to life: gather original evidence, showcase it in original ways and develop a signature form

Case study: HowTown. 15-plus-minute long-form videos that have clear chapter markers and distill key information while allowing users to easily go back or skip specific sections.

Mode 6: sensemake

 The vibe: looking for context to understand why a story matters and what to think about it
 The energy: exploring different viewpoints to form opinions. This mode is driven by curiosity and a desire for co-learning spaces
 The strategy: create spaces that invite perspective and transparency, helping audiences process disagreement and find orientation amid uncertainty

How to win at Sensemake mode:

• Create open conversations 
• Create open spaces for honest conversation where producers and audiences can think aloud
• Clearly disclose your intention, bias or viewpoint
• Orient and include the audience: vary show formats to deepen engagement and reach audiences across information needs
• Use livestreams to transform news into a shared, real-time act
• Use satire wisely (it should stimulate critical thinking): build recurring inside jokes to make complex topics feel approachable and foster a sense of belonging

Case Study: Morning Brew. Always linking to the original business story to anchor it to one company and how it’s being covered. Satire is created through the use of props, funny captions and over-the-top personas or characters.

---

3. The “Social” Mode: how news is shared

Mode 7: Socialize

 The vibe: sharing news to connect with others, express identity, or start a conversation. Sharing news is less a discrete consumption habit and more of a long-term relational action to be considered in its own context.
 The energy: there is no more valued, trusted source than friends and family, and often these connections help cut through content overwhelm.
 The strategy: build for sharing!

How to win the Social mode:

• Make sharing a “product feature” that makes the action effortless for readers
• Create content that people want to send to their peers
• Leading news producers tell human stories
• Consider using memes to condense meaning and inject humor

The takeaway for publishers

The “Next Gen” consumer is suffering from information overwhelm. To win, publishers must transition from being “content factories” to “mode-specific service providers.” Whether it’s winning the first two seconds of a Scroll or providing the deep-dive tools for Study, success in 2026 and beyond depends on meeting the audience exactly where they are.

The post Beyond the feed: mastering the 7 modes of next gen news engagement appeared first on Audiencers.

Khalil A. Cassimally is an audience consultant and coach who helps organisations align teams around change, user needs and AI. 

Audience and funnel work succeeds or fails on internal communication, yet few people get structured support to improve it. Drawing on recent research and his own practice, he shares a practical three-step way to use AI to:
> Understand your intervention style
> Identify strengths and gaps in how you communicate
> Take small, intentional steps to improve how your communication skills

Most audience and product work lives or dies on internal communication.

Yet, communication is also where many people struggle. Doing it well can be hard! And while senior executives may have access to coaching and leadership programmes, the people doing much of the day-to-day work rarely do.

This is where AI becomes interesting.

In an article published by Harvard Business Review (HBR) last year, researchers Katharina Lange and José Parra-Moyano explored whether AI could support intervention improvement in a way comparable to human coaching. Working with 167 global executives, they used AI to analyse real conversations and provide feedback, which participants then compared with feedback from human observers.

The results were telling. About 30% of participants received feedback that largely validated what they already believed about their intervention styles. More importantly, around 55% landed in what the researchers called the “zone of learning”: the feedback was both surprising and useful, sparking new insights.

My takeaway from this research isn’t that AI replaces human coaches but that reflective feedback can be made far more accessible than traditional coaching ever allows.

Inspired by their research, I built a repeatable process for using AI chatbots like ChatGPT and Gemini to improve my own interventions.

The research-informed process breaks down improvement into three phases:

• Understanding intervention style
• Identifying strengths and gaps
• Taking steps to improve how to intervene

Phase 1: understand your intervention style

Understanding how you currently communicate is the first step to improvement. To achieve this, you have to understand your preferred style – which you can do systematically with some anonymised transcripts and specific prompts.

> Use a shared language: John Heron’s categories of intervention

Most of us have default patterns of intervention. We need to name those patterns otherwise “communicating better” remains vague. To do this well, we need a shared language.

The post Improve how you communicate with colleagues: an AI method tested by researchers appeared first on Audiencers.