Times of India exec predicts traffic shift from volume to depth

A senior product leader at Times Internet outlined on December 15, 2024, how large language model technology will fundamentally reshape which journalism receives traffic from Google's platforms, predicting that publishers relying on volume strategies face declining referrals while those investing in original reporting gain advantage.

Ritvvij Parrikh, Senior Director of Product Management for AI at Times Internet, published analysis through Harvard's Nieman Journalism Lab forecasting that "the total click pool will shrink" while "remaining clicks will favor stories too semantically deep to be easily summarized" by artificial intelligence systems.

The prediction comes as Google has expanded AI Overviews to more than 100 countries throughout 2024 and 2025, with multiple studies documenting click-through rate decreases ranging from 34.5% to 54.6% for websites when AI summaries appear in search results.

Parrikh manages AI systems for The Times of India that serve audiences "without humans in the loop," according to his LinkedIn profile, requiring accuracy standards that inform his perspective on how retrieval technology affects publisher traffic. His LinkedIn post on December 18, 2024, emphasized that publishers must "dominate deep topics with fewer stories that are semantically deep (not news-peg based), contain significant net-new information relative to competitors, and are in formats that are not AI-readable."

The Times of India executive's analysis identifies Google's MUVERA system as evidence that the search company now represents stories as "semantic concepts (multi-vectors)" while searching for them cheaply. According to the Nieman Lab article, Discover's AI summary cards introduced in mid-2025 mainstream this approach by "grouping stories into topics, summarizing facts, and picking a single 'best-in-topic' story to showcase."

Traditional volume journalism strategies emerged because Google rewarded quantity, according to Parrikh's analysis. Publishers "chased a few trends, often drifting off-brand for traffic" using an approach of publishing many "good enough" stories targeting single keywords, where "a few hits subsidized the misses" and volume signaled authority.

Improved retrieval technology strains this model. Parrikh predicts six specific ecosystem developments as trends continue.

First, topic clustering will go mainstream across Search, Gemini, Discover, and YouTube rather than remaining restricted to niche products like Google News. "To cut LLM-generated clutter, expensive features like topic clustering, once restricted to niche products like Google News (since the 2010s), will go mainstream," according to the prediction.

Second, traffic will concentrate on fewer, deeper stories across Discover, YouTube, and Gemini platforms. Volume journalism will lose traffic as AI Overviews summarize shallow content. The remaining clicks "will favor stories too semantically deep to be easily summarized."

Third, most traffic will flow to "best-in-topic" stories containing competitively more net-new information. Google will filter by geography before identifying top stories "across publishers using algorithms like Information Gain (a 2022 Google patent) to reward unique data."

This competitive dynamic creates what Parrikh terms "verification latency," where retrieval augmented generation models view net-new information without consensus as potential hallucinations. Brand reputation and domain authority become "trust keys" allowing high-authority domains to bypass consensus filters while others wait for validation.

Fourth, large media organizations will invest across diverse topics, particularly in defensible niches where they can become canonical sources. Publishers will "spread their bets, aiming to become the 'canonical source' in specific topics in a certain language and geography to dominate the 'best-in-topic' selection."

Publishers will need to compete through defensible news-gathering including "feet on the ground (reportage); experts contributing knowledge, not just opinions; data journalism; analysis that is more sophisticated than what reasoning models can replicate," according to the analysis.

Fifth, Google Search's lucrative Top Stories carousel will continue prioritizing shallow recent stories for developing news while Discover, YouTube, and Gemini prefer fewer, deeper stories once news cycles settle. Parrikh characterizes this as "Fast Feed" for breaking news versus "Deep Feed" for settled topics.

Sixth, product teams will pivot before editorial operations. Current workflow, incentives, and content management system architecture favor "publish and forget" rather than "publish and maintain." Product teams will test auto-updating stories using large language models to de-duplicate overlapping articles "so on-platform users don't have to find information across dozens of near-identical articles."

The programmatic display model depends on volume and refresh rates while retrieval models move toward depth, creating monetization challenges. Five users bouncing through two shallow pages generating 5-10 impressions often produce more revenue than one user reading deeply for 10 minutes with 2-3 impressions, according to Parrikh's analysis.

Publishers must either pivot to reader revenue and affiliate models or prove to direct-sold advertisers that "one engaged reader is worth more than five users who bounce off easily."

Narrative strategy will bifurcate toward either radical voice or radical utility. High-voice approaches like The Atlantic will focus on "distinct narrative craft" where "the experience of reading is the product, not just the facts" with the goal of "irreducibility" making AI summaries feel like hollow substitutes. High-utility approaches like Axios's Smart Brevity optimize "time to value" through standardized production formats that improve machine legibility, potentially making publishers the preferred "structured data source" for AI models to cite.

Publishers attempting both approaches under one domain "will confuse the ranking algorithms and the users," according to Parrikh. He built an AI editor as a style linter to enforce these distinctions by "rejecting drafts that are too generic to be voiced but too unstructured to be useful."

The analysis received significant engagement on LinkedIn, where Ricky Sutton characterized Google as having "corrupted the economics of premium publishing by co-opting discovery, distribution and monetisation for two decades." Sutton argued that publishers "lazily ceded discovery to search and then to social; gave up distribution to Chrome and Meta; and then advertising to GAM and AdX."

According to Sutton's December 18 LinkedIn comment, Google told newsrooms to "drive volume to be visible in search" because "Google is an ad company" where "the more content it has, the more searches it gets, and the more the content is groomed for discovery, the more Google earns."

Research documented throughout 2025 supports concerns about traffic redistribution. Google Web Search traffic to news publishers declined from 51% to 27% between 2023 and 2025 while Discover feed climbed to 68%, according to NewzDash analysis of over 400 publishers worldwide announced December 23, 2025.

Traditional Google Search traffic dropped from approximately 16% to 10% of total referrals during the period when AI Overviews rolled out to more than 100 countries, with the inflection point occurring in late October 2024.

Research from Ahrefs examining 300,000 keywords found that AI-generated summaries reduce organic clicks by 34.5%when present in search results, comparing clickthrough rates for top positions with and without AI Overviews across identical time periods from March 2024 to March 2025.

Google executives have disputed traffic impact claims. Liz Reid, VP and Head of Google Search, announced on August 6, 2025, that "total organic click volume from Google Search to websites remains relatively stable year-over-year" while "average click quality has increased."

Nick Fox, Google's SVP of Knowledge and Information, told publishers on December 15, 2025, that optimizing for artificial intelligence search "is very similar, I would say, the same as how to perform well in traditional search," according to his appearance on the AI Inside podcast. Fox characterized third-party research showing traffic declines as "one-offs" and "cherry-picked analyses."

Publishers report dramatically different experiences. According to Bloomberg reporting published in April 2025, Google's introduction of AI-generated answers and algorithm changes caused website traffic to plummet for many publishers, with some reporting declines of 70% or more based on interviews with 25 publishers and data from Similarweb.

Mike Hardaker, founder of Mountain Weekly News which reviews outdoor gear, told Google's chief search scientist Pandu Nayak during an October 2024 meeting that he made $250,000 in gross revenue during 2023 but was relying on a food bank by the meeting date, according to detailed notes reviewed by Bloomberg. Nayak apologized but said "it wouldn't be reasonable for him to make any guarantees to the publishers because he couldn't predict what would happen."

The European Commission launched a formal antitrust investigation on December 9, 2025, examining whether Google violated EU competition rules by using publisher content for AI purposes without appropriate compensation or viable opt-out mechanisms.

Parrikh's prediction builds on his experience managing personalization systems that doubled engagement for Times of India while working with the Editor-in-Chief to redo information architecture that increased recirculation click-through rates by 95%. He also launched a subscription product achieving 120,000 subscribers in year one for the publication.

His analysis concludes that "in 2026, human-written volume journalism will have diminishing returns" as publishers must bifurcate business between a Fast Feed capturing breaking intent on Search and a Deep Feed for Discover, YouTube, and Gemini "where competitive 'information gain' wins."

The broader industry context shows fundamental shifts in search interaction patterns. Martin Splitt from Google addressed "The Great Decoupling" at the Google Search Central Live 2025 event in Warsaw on June 17, 2025, marking the first time a Google representative directly explained the separation of search impressions from clicks that publishers experienced since AI Overviews became standard.

Websites now receive significantly more impressions while experiencing substantial click decreases. Splitt suggested potential compensation through improved conversion quality, stating that "the clicks you get through search are more likely to be clicks that convert."

Research published by Semrush on June 9, 2025, found that AI search visitors demonstrate 4.4 times higher value compared to traditional organic search visitors when measured by conversion rates, though overall traffic volumes face temporary decline before reaching equilibrium.

The technological foundation for these changes involves semantic analysis capabilities. Screaming Frog announced version 22.0 of its SEO Spider tool on June 10, 2025, introducing semantic similarity analysis using large language model embeddings to identify duplicate and off-topic content, moving beyond traditional text matching methods.

Google's John Mueller issued warnings against certain automated content strategies. On August 27, 2025, Mueller stated that using large language models to build topic clusters creates "liability" and provides "reasons not to visit any part of your site," according to his comments targeting the widespread practice of using artificial intelligence to create comprehensive topic clusters.

For publishers, the analysis suggests that survival depends on either developing irreducible narrative voices that AI cannot satisfactorily summarize or creating highly structured utility content that becomes the canonical data source for AI systems to cite. The middle ground of generic, moderately detailed content faces elimination as AI summaries satisfy user needs without requiring clicks to publisher sites.

Timeline

• January 2025: Google deploys SearchGuard protections blocking automated scraping access to search results
• March 2025: AI Overview presence doubled by 116% following Core update, correlating with "crocodile mouth" traffic pattern emergence
• April 2025: Bloomberg reports website traffic plummeting 70% or more for multiple publishers after AI Overviews rollout
• May 21, 2025: Nick Fox defends AI search features on AI Inside podcast, questioning traffic decline studies
• June 9, 2025: Semrush study shows AI search visitors worth 4.4x more than traditional organic traffic by conversion rates
• June 10, 2025: Screaming Frog releases semantic similarity analysis using LLM embeddings for content clustering
• June 17, 2025: Martin Splitt explains "Great Decoupling" of impressions from clicks at Warsaw event
• August 6, 2025: Liz Reid claims click quality increased despite third-party research showing 34.5% decline
• August 27, 2025: John Mueller warns against LLM-generated topic clusters as building site liability
• December 9, 2025: European Commission launches antitrust investigation into Google's AI content usage
• December 13-15, 2025: Google core update destroys Discover traffic with some publishers reporting 98% drops
• December 15, 2024: Ritvvij Parrikh publishes predictions through Nieman Lab
• December 15, 2025: Nick Fox tells publishers AI search optimization unchanged from traditional SEO
• December 18, 2024: Parrikh shares analysis on LinkedIn emphasizing semantic depth requirements
• December 23, 2025: NewzDash analysis shows search traffic declined from 51% to 27% for news publishers between 2023-2025

Summary

Who: Ritvvij Parrikh, Senior Director of Product Management for AI at Times Internet, published analysis predicting fundamental changes to journalism traffic distribution from Google platforms.

What: The prediction outlines how semantic search technology and AI Overviews will shift traffic from volume journalism toward fewer, deeper stories containing unique information that AI systems cannot easily summarize. Publishers face bifurcation between irreducible narrative voices and structured utility content.

When: Published December 15, 2024, through Harvard's Nieman Journalism Lab as part of the Predictions for Journalism 2026 series, with supporting LinkedIn analysis posted December 18, 2024.

Where: The analysis addresses global impacts on publishers as Google expands AI Overviews across more than 100 countries, with specific focus on search ecosystem changes affecting Discover, YouTube, Gemini, and traditional Search platforms.

Why: Improved retrieval technology using multi-vector semantic representation strains the volume journalism model that Google previously rewarded. AI systems can cheaply search and summarize shallow content, reducing incentives for users to click through to websites unless content contains substantial net-new information, original reporting, or irreducible narrative craft that summaries cannot adequately convey.