Re-thinking Market Infrastructure for the Age of Narrative Data

By Xiaoqin Zhao, Co-Founder / CTO and Anna Tsiganchuk, Co-Founder / CEO at Aleta Index

Financial markets have always depended on information infrastructure. Exchanges, price feeds, and market-data vendors exist to ensure that participants receive reliable signals for decision-making. But the nature of the information feeding those decisions has changed dramatically.

Today, a growing share of market-relevant signals originates outside traditional financial data sources: news articles, social media discussions, blogs, and increasingly machine-generated content. This “alternative data” ecosystem has expanded rapidly over the past decade, promising deeper insights into sentiment, behaviour, and emerging risk.

Yet the approaches used to interpret this information have not evolved at the same pace.

When more data reduces clarity

Digital content is expanding rapidly. Global data volumes continue to grow by roughly 25%annually, according to Statista. In theory, more information should improve market efficiency by helping investors identify signals faster and price assets more accurately.

In practice, the opposite often occurs.

For many analysts and portfolio managers, the challenge is no longer access to information but filtering it. Decision-makers must assess thousands of articles, posts, and data points each day while determining which narratives genuinely influence markets and which represent short-lived noise.

This challenge has intensified with the rapid development of generative AI. The cost and speed of producing plausible news-style content has fallen sharply. Research from George Washington University identified a more than ten-fold increase in websites publishing AI-generated misinformation between 2022 and 2023. The result is an information environment where the credibility, influence, and relevance of content are increasingly difficult to evaluate in real time.

Markets move on narratives, not just facts

Traditional approaches to information verification treat content as isolated items. Fact-checking tools, source credibility rankings, and media reputation scores attempt to determine whether a claim is accurate or whether a source is trustworthy.

However, financial markets rarely respond to individual claims in isolation.

Instead, markets often move in response to narratives, clusters of related stories that gain momentum across multiple outlets and social channels. A narrative may combine partial facts, speculation, commentary, and sentiment. What matters is not only whether information is accurate, but how widely it spreads, who amplifies it, and when it reaches critical visibility.

This dynamic is particularly visible during earnings announcements, geopolitical events, or periods of market volatility. Competing narratives can emerge simultaneously, each shaping expectations in different ways.

Understanding which narrative gains traction, and how it propagates through the information ecosystem, has therefore become a key challenge for market participants.

The infrastructure gap

Despite the growth of alternative data providers, much of the underlying infrastructure still treats information as static text streams or sentiment scores. Articles are analysed individually, sources are ranked using fixed credibility metrics, and signals are aggregated through relatively simple models.

This approach struggles in an environment characterised by rapidly evolving narratives, coordinated amplification across platforms,

generative-AI content production, and shifting credibility signals.

The challenge is therefore not only analysing more data, but understanding how information behaves as a system.

Addressing this gap may require a new generation of market-data infrastructure focused on narrative intelligence rather than isolated documents.

From articles to narrative networks

Instead of evaluating each article independently, narrative-based systems organise large volumes of content into structures linking broader themes, daily topics, and individual reporting. Machine-learning models can then observe how these narratives propagate across sources, measuring patterns such as timing, sentiment alignment, and coordination between outlets.

Within this framework, influence emerges from patterns in how narratives spread and interact across the information landscape rather than from simple frequency counts or static reputation scores.

Source credibility can also be modelled dynamically, evolving as new evidence

appears. Instead of relying solely on historical authority, credibility signals may be  recalibrated based on how consistently a source participates in narratives that later correspond with market movements.

The result is a time-aware view of narrative formation and influence that may better reflect how markets process information.

An emerging layer of market infrastructure

As alternative data continues to expand, the next stage of market infrastructure may focus less on collecting information and more on understanding how it moves.

Several emerging solutions are exploring this direction. Aleta Index, for example, is developing infrastructure designed to track how claims and narratives propagate across sources over time, and whether those narratives correspond with market outcomes.

Within this approach, source credibility is treated as a dynamic process rather than a fixed attribute. Credibility evolves through observable patterns, such as whether claims  are later corroborated, whether they align with market outcomes, and whether a source consistently leads or follows in developing narratives.

The system can also surface competing interpretations within a single narrative. Different sources often advance conflicting views of the same events. Observing how these interpretations gain or lose traction, and which sources most reliably anticipate market direction, provides a more structured view of information influence than sentiment aggregation alone.

For financial institutions, this reflects a shift in how alternative data infrastructure is designed: from static source rankings to continuous observation of how information behaves, and from isolated document analysis to mapping the dynamics of narrative formation in real time.

The practical implication is a layer of intelligence that sits between raw news flows and investment decisions, not as a trading signal itself, but as a more reliable lens through which the credibility and relevance of emerging narratives can be assessed.