Avinash Mudaliar: The Attention Economy Is Dead. Welcome to the Habit Economy.

image avinash mudaliar mudaliar memo #1 mediabrief SGJGMudaliarMemo is a new weekly column for MediaBrief by Avinash Mudaliar — inventor of the Saregama Carvaan, co-founder of Gaana, and now CEO and Co-founder – HT labs. In this first piece, he writes:

The industry built a content firehose. The audience turned off the tap. Here is my diagnosis of what actually happened — and why the rarest asset in media is no longer content, but trust.

There is a moment every media leader has lived through — usually around midnight, staring at a dashboard that shows traffic going up and revenue going sideways — where the comfortable assumptions of the last twenty years begin to quietly fall apart.

We built this industry on a simple thesis: more content, more reach, more formats, more distribution. Scale the supply; capture the demand. It made sense for a world where information was scarce, platforms were neutral pipes, and search reliably sent curious people to the places that had answers. We were the library in a town with no library. Of course people would come.

That town no longer exists.

The Wrong Thesis

What replaced it is an environment of such radical content abundance that the cognitive burden of simply choosing has become the primary friction in the user’s day. Not finding. Not quality. Choosing. Gloria Mark at UC Irvine has spent two decades measuring the collapse in sustained attention: from roughly 150 seconds of average focused screen-task time in 2004, to 75 seconds by 2012, to 47 seconds by 2024. Nielsen Norman’s 2026 research tracks a further decline to 43 seconds, with 566 context-switches in a single working day — approximately one switch every 51 seconds.

We are designing for audiences whose attention fragments every minute, and wondering why our longform is underperforming.

The supply problem has a human cost. The Reuters Institute’s Digital News Report 2025, which surveyed audiences across more than 40 countries, found that 40 percent of news consumers sometimes or often actively avoid the news — up from 29 percent in 2017. Among 18 to 24-year-olds, 43 percent say they feel exhausted by the volume of news; 31 percent feel overwhelmed by it. We built a firehose. They turned off the tap.

What the Data Actually Shows

The products that are winning the daily relationship with users are not doing so through volume. They are doing so through something much harder to manufacture: habit.

TikTok users spend an average of 95 minutes per day on the platform across 19 sessions — sessions that are nearly four times longer than Instagram’s. Not because TikTok’s content is uniquely excellent. Because TikTok has engineered the near-complete elimination of decision fatigue. The feed decides. The user receives. The friction evaporates.
Source: Sensor Tower / Financial Times synthesis, 2025

Spotify sits at roughly 2 percent monthly churn. Video streaming platforms typically sit at 5 to 10 percent. The difference is not catalogue size. It is habitual, ambient listening. When Spotify introduced its AI DJ feature, users who adopted it averaged 140 minutes of listening per day versus 99 minutes for non-AI users — a 41 percent engagement lift achieved not by adding more music, but by removing one more micro-decision.
Source: Spotify earnings, 2024–25; Churnkey, 2025

McKinsey’s 2025 Attention Equation research (7,000 global respondents) found that a 10 percent increase in average user focus is associated with a 17 percent increase in consumer spending. Focus times intent equals value. Everything else is impressions theatre.
Source: McKinsey & Company, 2025

These are not outliers. They are the clearest available signal about what the next chapter of digital media looks like. Habit is not a nice-to-have. It is the only sustainable competitive position.

The AI Complication

Into this environment of fragmented attention and eroding habit, generative AI has introduced a development that will shape the next decade more than any other: the industrialisation of content production.

By late 2024, according to analysis of more than 65,000 English-language articles, over 50 percent of new web content was AI-generated. AWS researchers estimate that 57 percent of online content is now created or translated by AI. We have crossed a threshold: the majority of text on the internet is now machine-authored.

Audiences are beginning to sense it. The Reuters Institute’s Generative AI and News Report 2025 found that only 12 percent of users are comfortable with news made entirely by AI, versus 62 percent who are comfortable with fully human-made journalism. KPMG and Melbourne Business School’s global study — 48,000 respondents across 47 countries — found that 66 percent of people use AI regularly, but only 46 percent trust it. The trust gap is not closing. It is widening.

AI has made content infinite. It has also made trust finite.

The Discovery Problem

There is a structural shift happening beneath these cultural changes that is, arguably, the most consequential for anyone operating in publishing.

Google search referral traffic to publishers declined approximately 33 percent globally between November 2024 and November 2025, according to Chartbeat data cited in Reuters Institute’s Trends and Predictions 2026 report. In the United States, the decline was roughly 38 percent. When Google’s AI Overview appears in search results, only 8 percent of users click any link — compared to 15 percent without one. Publishers surveyed expect search referrals to fall 43 percent within three years. A fifth anticipate losing more than 75 percent.

Search was the single largest source of referral traffic for most publishers for nearly two decades. That model is deteriorating rapidly. AI assistants are emerging as alternative discovery surfaces, but they currently deliver a fraction of the referral value — Google still delivers roughly 500 times more publisher referrals than ChatGPT. We are in the gap between an old model that is eroding and a new one that has not yet stabilised.

My Diagnosis

The core problem is this: we built organisations optimised for a world that no longer exists. We built for discovery through search and social, measured ourselves in page views and unique visitors, monetised through advertising inventory, and treated the audience as an aggregate rather than a relationship.

That worked when information was scarce, search was reliable, social platforms were open pipes, and attention was relatively abundant. None of those conditions still fully hold.

What the data now clearly supports is that the competitive advantage in publishing for the next decade will not belong to the organisations with the most content. It will belong to the organisations that have built the most durable habits.

A page view is an event. Attention is a relationship. Trust is an asset. The organisations building actual habits — the push notification users check before coffee, the newsletter that explains the week, the product review saved when considering a purchase, the food destination that shapes what families cook — are building something that neither search algorithms nor AI assistants can easily replicate.

Friction is the new churn. Habit is the new reach.

How I Am Now Looking at Things

When I evaluate any decision about content, product, or platform now, the first question is no longer ‘how much traffic will this generate?’ It is: does this create a reason to come back?

The second question is: what relationship does this build?

The third is: what does the user do next?

That last question represents the most important shift. The most sophisticated media businesses are beginning to understand that their job is not simply to inform — it is to help users navigate from information to understanding to decision. That is a fundamentally different editorial mission, and a fundamentally more defensible commercial one.

AI makes information abundant. What it does not make abundant is judgement, context, trust, and the specific human knowledge that comes from having spent years watching a market, understanding its players, testing its products, and knowing which numbers not to believe.

In a world flooded with machine-generated summaries, the scarce assets are exactly the ones journalism has always produced: original reporting, expert testing, earned trust, and the accumulated knowledge of people who have been paying attention for a long time. AI made those assets more valuable, not less.

Publishers that send even a single push notification in the first 90 days of a user relationship see 147 percent higher long-term retention. Those sending weekly push notifications see six times higher 90-day retention than those that don’t. The audience that stays, returns, saves, shares, and acts is worth infinitely more than the audience that arrives from search and immediately leaves.

The publishers that survive the next decade will not be the biggest. They will be the most habit-forming — and the ones that understand that earning a daily appointment is a more defensible position than winning a search query.

AI multiplied the supply of words. Trust is now the only scarce input in the system.


Avinash Mudaliar is CEO and co-founder of HT Labs. Earlier, he was CPO of Network 18;  he co-founded Gaana, invented the Saregama Carvaan, and was Editor-in-Chief of RAVE. His writing has appeared in The Times of India, Hindustan Times, OPEN, Firstpost, HuffPost and OTTplay. He writes MudaliarMemo, MediaBrief’s weekly column.