Social Media Algorithms Now Reward Watch Time, Not Conversation

Open Instagram, TikTok, YouTube or Facebook in 2026 and the math the platforms do for you has very little to do with what your friends posted. It has everything to do with how long you keep watching. Hootsuite’s 2026 ranking-signal guide documents the shift across every major network: AI-powered social media algorithms determine what users see every time they open a social media app an average of 141 minutes per day worldwide, and the algorithm’s job is to keep them there. The brands that built their social playbooks around comments, shares and community are now playing a game the platforms stopped scoring that way years ago.

The Algorithm Shift: From Conversation to Watch Time

The 2026 ranking-signal guide from Hootsuite walks through every major platform and the same signals keep rising to the top: watch time, predicted engagement, send rate. For Instagram, the guide quotes Adam Mosseri, Head of Instagram, naming the top three signals directly: watch time, likes, and sends. TikTok’s strongest signal is watch time. YouTube’s is watch time. Facebook’s is predicted engagement, with watch time built into the prediction. LinkedIn still weights comments and discussion quality, but only inside a model that increasingly predicts whether a viewer will scroll past the post in three seconds or stay.

Inside the agencies that buy social for a living, the new arithmetic is openly named. Aprajita Biala, National Planning Head at Cheil India, put it bluntly: a 200-comment carousel will lose distribution to a 2-million-view reel that drew 40 comments, because the reel holds attention longer. The platform is not rewarding conversation. It is rewarding the absence of an off-ramp.

The Participation Gap Was Always There

The lurker majority is not new. Jakob Nielsen’s 2006 article on participation inequality formalized what Will Hill at Bell Communications Research had been studying since the early 1990s. User participation online, Nielsen wrote, often more or less follows a 90-9-1 rule: 90% of users are lurkers, who read or observe but don’t contribute, 9% of users contribute from time to time but other priorities dominate their time, and 1% of users participate a lot and account for most contributions. The rule has held in every online community and multi-user service that has ever been studied.

Shan Jain, an independent consultant, draws the line that matters now: what changed isn’t user behaviour, it’s that platforms stopped pretending. Algorithmic feeds monetise watch time, not conversation. The 90% who never comment were always there. The platforms simply stopped pretending the other 9% were the audience they were designing for.

• 90% of online community users are lurkers (Jakob Nielsen, 2006)
• 9% contribute occasionally (Jakob Nielsen, 2006)
• 1% account for most contributions (Jakob Nielsen, 2006)
• 141 minutes per day spent on social apps worldwide (Hootsuite, 2026)

What Marketers Are Now Saying, Out Loud

Industry voices are catching up to the lurker reality. Biala, who leads planning at one of South Korea’s largest agency networks, framed comments and shares as algorithmic fuel rather than relationship proof: conversation hasn’t scaled but distribution has. Saurabh Munjal, Co-Founder and CEO of Lahori Zeera, took the argument into business terms. Most consumers don’t wake up wanting to have conversations with brands, he said, they simply want brands that consistently deliver value.

Jain pushed the broadcast framing further. Social media today is less like a neighbourhood cafe and more like an airport terminal, she said. Millions of people pass through it every day, but very few stop long enough to have a meaningful conversation. The platforms are still labelled social, but the design rewards throughput.

Budgets are following the rhetoric. NP Digital’s November 2024 survey of 11,093 marketers worldwide found that 58% of marketers planned to decrease their organic social budget. Twenty-five percent planned to increase it. Seventeen percent expected budgets to stay flat. Reductions outweigh increases by more than 2 to 1.

Discovery Is Moving to AI Intermediaries

The shift is not finished. Gartner’s February 2024 prediction on search volume, the most-cited primary forecast in this space, said that by 2026 traditional search engine volume will drop 25%, with search marketing losing market share to AI chatbots and other virtual agents. The 50% figure that has since circulated in industry decks is a derivative projection that does not trace to a primary Gartner release I could verify in this session, so the 25% number is the one this piece is built on.

Generative AI solutions are becoming substitute answer engines, replacing user queries that previously may have been executed in traditional search engines. This will force companies to rethink their marketing channels strategy as GenAI becomes more embedded across all aspects of the enterprise.

That is Alan Antin, Vice President Analyst at Gartner, on the record in the same release. The substitution is already visible in the data. Merkle’s February 2025 analysis of AI discoverability, written by Prachi Gururaj, reported that 60% of search queries now end without a click, that one in five retail purchasing decisions is directly influenced by AI, and that nearly six in ten shoppers have replaced traditional search with AI tools as their primary discovery method. The front door to any given brand has moved, and for many organizations, it has moved without them.

The new front door has two audiences. The same Merkle piece names the dual-audience problem plainly: digital experiences must serve two fundamentally different audiences simultaneously. The human buyer wants immersive, brand-rich experiences. The AI agent needs structured, machine-readable data to accurately represent products and services. Brands that built their web presence for the human buyer alone are now invisible to the buyer that increasingly decides what the human buyer ever sees.

What Has to Change in Measurement

Shikha Davessar, Managing Partner at the agency 22feet, named the diagnosis most marketers still avoid: the honest answer for most brands is that they’ve mistaken platform presence for relationship ownership. Munjal, of Lahori Zeera, took the case into the metrics: engagement has become an increasingly misleading metric. A post with millions of views may have little commercial impact, while a brand with relatively modest engagement can enjoy exceptional repeat purchase. Virality, he added, is not a business model.

Biala’s layered model maps cleanly to what is actually measurable inside and outside a platform. The point is not to stop measuring the top of the funnel. The point is to stop treating the top of the funnel as the funnel.

• Views, saves, and watch-throughs (presence at the top of the funnel)
• Link clicks, profile visits, and branded search lift (intent signals)
• Repeat engagement and owned-base growth (compounding, business-level metrics)
• Penetration, repeat consumption, share of wallet, and mental availability (per Munjal’s framework)

Munjal’s framework is the one that survives an algorithm change. Penetration, repeat consumption, and mental availability are the numbers that move on owned channels, in retail, and in product experience, not in a platform’s ranking model.

Owned Channels Are Where the Relationship Lives

Biala drew the line that turns the whole debate from a media question into a property question. Social is where you get found and where you transact, she said, but this is rented infrastructure with someone else’s data and algorithm. Owned assets, CRM databases, loyalty programmes, mobile apps, retail ecosystems, sit outside the platform’s reach. A policy change, an algorithm shift, or a sunset product cannot touch them.

Jain offered the stress test every brand should be willing to run on itself: if Instagram disappeared tomorrow, how much of your customer relationship would survive? If the answer is not much, then the algorithm owns the audience, not the brand. In her reading, the things that count as owned in a market like India are dealer networks, WhatsApp communities, retail relationships, and the kirana owner’s recommendation. Translated globally, the equivalent list is dealer networks, retailer relationships, owned communities, and first-party data, the assets a brand can address without asking a platform for permission. Rented reach is what most social followings actually are. Owned reach is what a CRM is for.

Munjal closed the arc with the line that ties the whole reckoning together: algorithms can decide who sees you today. They cannot decide whether consumers choose you tomorrow. The work between rented reach and durable preference is product consistency, retail experience, customer service, and the reputation a brand carries into the next discovery layer. That next layer, the AI intermediary, will judge brands on the same things.

Reading the platform data, the Gartner forecast, the Merkle analysis, and the agency voices together, the shape of the next decade of brand building is clearer than the spend on social would suggest. Social media’s job, on the evidence, is to get a brand noticed, not to win it customers. The work that wins customers happens in CRM depth, retail relationships, and the structured data an AI intermediary can read.

Frequently Asked Questions

Why do social media algorithms now prioritize watch time over engagement?

Because watch time is the metric that maps directly to ad revenue. Hootsuite’s 2026 ranking-signal guide and Adam Mosseri’s own statements on Instagram both name watch time, alongside likes and sends, as the top three ranking signals for the platform in 2026. The platforms have stopped pretending the goal is conversation; the goal is the next minute of viewing.

What is the 90-9-1 rule, and is it still true?

The 90-9-1 rule, formalized by Jakob Nielsen in his 2006 article on participation inequality, holds that 90% of online community users are lurkers, 9% contribute occasionally, and 1% account for most content. Nielsen’s own assessment is that the rule has held in every online community and multi-user service ever studied. Modern platforms have not abolished it; they have built ranking systems around it.

Are brands still investing in organic social media?

No, not at the same rate. NP Digital’s November 2024 survey of 11,093 marketers worldwide found 58% planned to cut their organic social budget, against 25% planning increases and 17% expecting budgets to stay flat. Budget reductions outweigh increases by more than 2 to 1.

What is AI discoverability, and why does it matter for brands?

AI discoverability is the practice of structuring a brand’s digital presence so AI agents can interpret and recommend it. Gartner’s February 2024 release forecast a 25% drop in traditional search volume by 2026 as chatbots absorb queries. Merkle’s February 2025 analysis reported that 60% of searches now end without a click and that nearly six in ten shoppers have replaced traditional search with AI tools as their primary discovery method.

What counts as an owned channel for a brand?

Owned channels are the assets a brand controls directly: CRM databases, loyalty programmes, mobile apps, retail and dealer relationships, and first-party data. Unlike a platform feed, none of these is subject to a sudden algorithm change. The agency voices in this piece frame them as the only infrastructure that survives an Instagram, a policy change, or an AI intermediary.

Read next: a novelist’s reckoning with her own seven-hour weekly scroll traces the same shift through one writer’s calendar.