Beyond the Public Feed

The Rise of Private Sharing on Social Networks

Iceberg illustrating public vs private sharing

Public Spaces

(e.g., public pages, groups)

Private Spaces

(private groups, direct messages)

Understanding content dynamics requires examining both spaces.

The Future is Private

Forbes article: Mark Zuckerberg Says The Future Is Encrypted, Private Messaging

Platforms themselves are pivoting towards private, encrypted communication, acknowledging a massive shift in user behavior.

The Research Gap

While we know this shift is happening, there has been no quantitative assessment of this "spillover". How much of the conversation are we missing by only looking at public data? This study is the first to measure it.

Why Quantify Private Sharing?

An incomplete picture of the information ecosystem leads to flawed conclusions.

1. The Blind Spot in Misinformation Research

This YouTube video spreading election misinformation shows nearly 1 million views. Standard research tools would see this high engagement.

But on Facebook, public data tools like CrowdTangle show only ~2,800 interactions. The other ~109,000 interactions happened privately, hidden from researchers.

CrowdTangle data showing low public shares

Without accounting for private sharing, we would underestimate this link's Facebook reach by over 97%.

2. Measuring True Content Impact

News outlets, creators, and marketers cannot gauge the true reach of their content if a majority of its shares are invisible. Public metrics like 'likes' or 'shares' on a page become vanity metrics if the bulk of dissemination and discussion happens in private DMs and groups.

3. Understanding Platform Dynamics

Why do people share certain content (e.g., controversial topics, niche interests) privately? Understanding these motivations is crucial for platforms, policymakers, and researchers to grasp how online social behavior and discourse are evolving away from public town squares.

A Quantitative Approach

Using Meta's CrowdTangle to measure public and private shares.

Use the CrowdTangle Chrome Extension

The CrowdTangle Chrome Extension shows public interactions for a URL from Facebook Pages, public groups, and verified profiles.

CrowdTangle extension showing public shares

Get Public and Total Shares

The CrowdTangle Link Checker tool shows all Facebook activity, including public shares AND private shares (in DMs, private groups, etc.)

CrowdTangle Link Checker showing total interactions

The Key Metric: Share Ratio

By combining these two data points, the study calculates:

Ratio = Public Shares / Total Shares

A Ratio near 0 means mostly private sharing. A Ratio near 1 means mostly public sharing.

A Diverse and Comprehensive Dataset

The study analyzed URLs from a wide variety of sources to get a holistic view.

Mainstream News

NYT, Fox News

Biased News

Breitbart, Newsmax

Low-Quality Sources

Misinformation sites

Fact-Checking

Snopes, PolitiFact

Social Platforms

YouTube, Reddit, WhatsApp

Alternative Media

Substack

COVID-19 News

Large topical dataset

Category Analysis

Huffington Post data

Key Results & Findings

Sharing behavior varies dramatically by source, topic, and category.

1. Mainstream vs. Biased & Fringe News

Key Takeaways

The chart shows the fraction of a source's URLs (y-axis) that have a public-to-total share ratio less than a certain value (x-axis).

Steep curves (NYTimes, Fox News): Indicate content is overwhelmingly private. For NYTimes, ~72% of its URLs have a public share ratio of 0.2 or less.
Shallow curves (Breitbart, Newsmax): Indicate content is shared more publicly. Newsmax, the shallowest curve, requires a ratio of ~0.7 to cover 70% of its URLs.
COVID News (in-between): Shows a mix of sharing behaviors, landing between the highly private mainstream news and more public biased sources.

Click labels on the chart to toggle sources!

2. Sharing Varies by News Category

This chart shows the extremes: the 5 most privately-shared and 5 most publicly-shared categories.

...

28 Other Categories

...

3. Controversial Topics Go Dark

Topic modeling of COVID-19 news reveals a concerning trend.

Implications for Research

The findings challenge conventional research methods and highlight critical gaps.

Public Data is an Incomplete Picture

Studies focusing only on public data are missing the majority of sharing activity for many content types (e.g., 80%+ of mainstream news shares). This can lead to biased results and a flawed understanding of content reach and influence.

Misinformation is Underestimated

The high level of private sharing for low-quality and conspiratorial content means its true prevalence is hidden. This poses a major obstacle for researchers and fact-checkers trying to quantify and combat misinformation.

A Call for New Methods

Since platform data access is limited, the paper calls for researchers to explore alternative models like privacy-preserving data donation and independent qualitative studies to better understand the dynamics of private sharing.