Investor Alert
Jurica Dujmovic

Your Digital Self Archives | Email alerts

May 19, 2022, 12:21 p.m. EDT

Why it’s so hard to know how big Twitter’s bot problem is

Watchlist Relevance

Want to see how this story relates to your watchlist?

Just add items to create a watchlist now:

or Cancel Already have a watchlist? Log In

By Jurica Dujmovic

The dust hasn’t settled on the bombshell news that Tesla CEO Elon Musk aims to buy Twitter. Among the hubbub, a new issue has arisen.

According to this MarketWatch piece, “[A]s stock prices have plunged in the overall market downturn, Musk has obviously suffered from buyer’s remorse and is saying that the deal is ‘on hold.’”

While the piece claims that Musk has halted the acquisition because he got cold feet — which might be true — his public reasoning behind the move is what I want to focus on today: Twitter’s bot problem.

Musk has been vocal about the issue since he made a Twitter account in 2010, and rightly so: Thousands of bots and scammers pose as the billionaire, luring the naïve and uninitiated into sending them crypto and otherwise tricking them. And these bots are just the tip of the iceberg.

Governments, political parties, companies and hackers all have their own bots, and these fake accounts have found a way to circumvent authentication protocols and inject themselves into the network.

The result is not only the rise in cybercrime, but also rampant misinformation, harassment and artificial amplification of certain threads and tweets via fake engagement of multiple accounts.

Needless to say, Twitter bots are a problem that needs to be solved, but how big is it?

According to Twitter CEO Parag Agrawal, bots comprise around 5% of accounts, while Musk claims the number could be well over 90%.

While the company published an interesting report about detecting and banning malicious accounts, and Agrawal made a thread about it as well, the fact remains that the process isn’t perfect, and hasn’t been for a while now.

In 2014, Twitter acknowledged that 8.5% of its userbase, which at that time was 23 million accounts, were bots. A study published in 2017 claimed that number rose to roughly 47 million or 15% of accounts. Furthermore, between October 2018 and December 2021, Twitter purged 200 million tweets and 9 terabytes of media, claiming them to be “platform manipulation campaigns originating from 17 countries.”

In 2021, the company actually verified several bot accounts, and deleted them only after a warning posted by a data scientist Conspirador Norteño .

Judging by this, it seems that the Twitter bot problem is far greater than Agrawal would have you believe. And I understand why he would go out of his way to dispel concerns. For one, a bot-ridden platform is unappealing to advertisers.

Not only is advertising money wasted on fake accounts, but such engagement skews advertising metrics and can lead to wrong conclusions when optimizing advertising strategies and launching campaigns.

The nature of bots is irrelevant as far as advertisers are concerned — whether they’re spamming, posting disinformation or just standing there passively, they are not, and never will be, customers.

Having said that, there’s another alternative theory to consider: What if both Musk and Agrawal were correct in their assumptions? What if around 5% of all users on Twitter are harmful bots, while a big chunk of the total number are passive bot accounts, including those made by Twitter and other companies for the simple purpose of artificially inflating the userbase?

Twitter’s own API allows bot creation, so it’s not beyond the realm of possibility that the platform would use the same tools to blow up the number of users. This certainly make sense when taking into account the company’s aggressive growth plans for 2023 .

If this is true, then uncovering this scheme could considerably dilute Twitter’s value for advertisers and cause them to cease putting the money into the platform.

For a company that hasn’t been profitable for most of its existence, this would be a big problem.

My colleagues at other outlets claim that Musk doesn’t provide valid proof for his 90% claim, but the same can be said about Twitter’s CEO. His “transparency” and “facts” may be nothing but fabrications, and no one will know for sure until the platform’s code goes open source and processes are made truly transparent.

All we have right now is speculation. Twitter claims it’s doing enough to curb the bot numbers. Musk says there’s more work to be done.

What do you think? Have you ever run into Twitter bots during your time on the platform? Let me know in the comments section below.

This Story has 0 Comments
Be the first to comment
More News In

Story Conversation

Commenting FAQs »

Partner Center

Link to MarketWatch's Slice.