The Art of the Shadowdan

Web3 social is supposed to be different than Web2 social; instead, it’s more of the same

Geoff Golberg
7 min readMay 9, 2024

The term “Web3” can take on many meanings, depending who you ask.

To me, Web3 is the next iteration of the internet, where individual users will regain power that has been amassed by Big Tech companies.

As the saying goes, “if you’re not paying for the product, you are the product.” And within the walled gardens of Big Tech, users have very little control over their data and content.

Building audience requires consistent content creation, and generally can be a time consuming process. With Web2, it’s important to keep in mind that you’re building on someone else’s land. Big Tech has so much power, they can opaquely and unaccountably remove someone from their platforms, for example.

I know, as I have been deplatformed not once, but twice. In fact, that experience is what has fueled my interest in blockchain technology, and its potential to subvert the tech monopolies. In both instances, I was rugged as a direct result of my attempts at holding Big Tech to account:

Whereas deplatforming involves the outright removal of an account, shadowbanning involves restricting an account. Shadowbanning may actually be more sinister than deplatforming, as the user is generally not informed of the occurrence — they are simply left with uncertainty and a loss of engagement.

Enter Farcaster

Farcaster is a decentralized social protocol, one of a handful, where developers are building out various social networks — ones that are based on blockchain technology.

In the case of Farcaster, the protocol’s largest client/application is Warpcast, which offers a Twitter-like user experience. Farcaster — and Warpcast, which is developed by the same team as the protocol — are products from Merkle Manufactory, a start-up that was valued at a billion dollars as of their most recent capital raise in March.

Key benefits of Web3 social (i.e. decentralized social protocols) when compared to Big Tech (Web2 social) include:

  1. Social graph portability/ownership
  2. Censorship resistance
  3. Decentralization
  4. Permissionless nature

In theory, both deplatforming (censorship resistance) and shadowbanning (transparent nature of blockchain data & availability of off-chain data) have a much lower likelihood of taking place in Web3 social when compared to Web2 social.

Warpcast’s “Power Badge”

Warpcast uses a PageRank-like algorithm to assign Power Badge designations to a subset of their users.

One key benefit to having a Power Badge is increased visibility for your account (more likely to be recommended as an account to follow, ranking higher in search results). Another key benefit is the ability to send direct casts/messages to accounts that are not following your account.

According to the Warpcast blog, below are the “most important things you can do to improve your probability of getting a power badge:”

According to Warpcast’s CEO, Dan Romero, the Power Badge algorithm may be summarized as such:

TLDR: being active, and posting content that drives engagement (recasts, replies, likes) from other Power Badge-holders is what is required to secure a Power Badge

Don’t Trust, Verify

Last week, my Warpcast account had a Power Badge; this week, it does not.

Given last week was my best week on Warpcast (in terms of followers gained and engagement from other Power Badge-holders), I was most certainly surprised when I realized on Tuesday that my account no longer had a Power Badge:

It was when I went to send a direct cast/message to an account that I had been messaging with prior — but that was not following mine — that I realized something was up, as I was no longer able to message said account. Shortly thereafter, I noticed that a Power Badge no longer appeared next to my account’s display name.

This seemed very odd, given, as mentioned earlier, that I had the Power Badge last week (from my inbound engagement two weeks ago), and last week — used to determine which accounts receive Power Badges this week — was my best week yet.

Thankfully, all I needed data-wise to compare the performance of my account last week vs. the accounts (4,344) that received Power Badges this week is easily accessible.

So, that’s exactly what I did next, and here is what I found:

Recasts

I received a total of 101 recasts last week, 88 of which (87.1%) were from accounts that had Power Badges.

Of the 4,344 accounts that currently have Power Badges, a mere 81 (less than 2%) received more recasts from Power Badge-holding accounts last week than my account.

Moreover, 87.1% of my total recasts coming from Power Badge-holding accounts — among the 4,344 current Power Badge-holders that had at least 100 total recasts last week — is the highest percentage among the dataset (from the 545 accounts that fit said criteria):

Each account above has a Power Badge; mine does not

Such a large percentage of my recasts being linked to Power Badge-holding accounts is indicative of power users enjoying the content I shared.

Likes

I received a total of 494 likes last week, 359 of which (72.7%) were from accounts that had Power Badges.

Of the 4,344 accounts that currently have Power Badges, 349 (8%) received more likes from Power Badge-holding accounts last week than my account.

Moreover, 72.7% of my total likes coming from Power Badge-holding accounts — among the 4,344 current Power Badge-holders that had at least 100 total likes last week — is the 8th highest percentage among the dataset (from the 1,678 accounts that fit said criteria):

Each account above has a Power Badge; mine does not

Such a large percentage of my likes being linked to Power Badge-holding accounts is indicative of power users enjoying the content I shared.

Replies

I received a total of 129 replies last week, 109 of which (84.5%) were from accounts that had Power Badges.

Of the 4,344 accounts that currently have Power Badges, 451 (10.4%) received more replies from Power Badge-holding accounts last week than my account.

Moreover, 84.5% of my total replies coming from Power Badge-holding accounts — among the 4,344 current Power Badge-holders that had at least 100 total replies last week — is the 4th highest percentage among the dataset (from the 995 accounts that fit said criteria):

Each account above has a Power Badge; mine does not

Such a large percentage of my replies being linked to Power Badge-holding accounts is indicative of power users enjoying discussing the content I shared.

Accounts With No Activity

Of the 4,344 accounts that currently have Power Badges, 918 (21.1%) did not cast or reply to any casts last week.

The Power Badge algorithm explanation seems to imply that accounts must be active to earn a Power Badge, hence this is particularly odd. Given the lack of activity last week for those 918 accounts, it indicates there is a large set of accounts that are whitelisted to have Power Badges.

Those 918 accounts may be found here.

Conclusion

Web3 social when compared to Web2 social — among other things — differs in that the likelihood of deplatforming and shadowbanning should, in theory, be significantly reduced given censorship resistance and the transparent nature of blockchain data & availability of off-chain data.

Specific to Warpcast, losing a Power Badge effectively functions as shadowban, as it reduces an account’s visibility within the network.

And, in my case, the data simply doesn’t justify why my account would lose its Power Badge, while 918 accounts that were not active on Warpcast last week emerged this week with Power Badges. Moreover, my account’s engagement metrics last week — across recasts, likes and replies — fares incredibly well when compared to accounts that earned Power Badges this week.

Given I have been vocal in calling out Farcaster for presenting protocol metrics which are heavily inflated by accounts engaging in CIB (coordinated inauthentic behavior) and platform manipulation, more broadly, it certainly feels as though the Farcaster team is attempting to silence my voice by reducing the visibility of my account.

Further reducing my ability to get my message out, I am no longer able to send direct casts/messages to accounts that are not following mine — a tactic I have utilized heavily to connect with other users in the past.

And finally, my account is no longer able to cast videos — presumably a result of no longer having a Power Badge.

Web2 social’s shadowban has evolved to Web3 social’s shadowdan.

Data was sourced via The Indexing Company.

Social Forensics maps and monitors social connections and activity.

We create purposefully designed tools to manage social data analytics needs across various industries. Our focus is audience segmentation and identifying coordinated inauthentic behavior (CIB) across social media platforms.

Geoff Golberg is an NYC-based researcher (and entrepreneur) who is fascinated by graph visualization/network analysis — more specifically, when applied to social networks and blockchain activity. His experience spans structured finance, ad tech, and digital marketing/customer acquisition, both at startups and public companies.

Geoff is the Founder/CEO/Janitor of Social Forensics.

--

--

Geoff Golberg
Geoff Golberg

Written by Geoff Golberg

CEO & Founder, Social Forensics | Previously: Co-Founder, Elementus | Featured in BBC, CNN, BuzzFeed, and Quartz, among others | SocialForensics.com

No responses yet