Many of Twitch’s biggest personalities saw their viewcounts plummet last week as the streaming platform’s crackdown on viewbots seemingly took effect. Twitch had announced in late July that it would start cracking down on automated viewership.
On X, formerly known as Twitter, the popular streaming platform announced it had “meaningfully” implemented changes that would allow it to more effectively “identify viewbots, inauthentic viewership, and other potentially fake engagement.” It said the plan would be rolled out over following weeks.
Late last week, those changes started to be felt. On August 23, industry analyst Zach Bussey asserted that Twitch’s site-wide viewership was “down 5–22%” compared to a week prior, a claim contested by Twitch claiming third-party sources were citing “incorrect data”.
Although anecdotal, more in-depth analysis is likely to come out over the next few weeks or months. The data is somewhat skewed by the fact that some creators were noticeably absent from streaming between August 21–23.
By August 25, two days later, average view counts had returned to normal. Bussey suggested that Twitch had walked back its changes or that “every viewbot service has found a solution already.”
Combating artificial views isn’t anything new for the Amazon-owned platform, but a ramping up of intensity of ongoing efforts, as CEO Dan Clancy said in an interview with Noah Kara on July 31.
What is Viewbotting?
Viewbotting is the act of artificially inflating the view count of online content, usually streams and videos. This can take the form of web automation with bots and scripts, but can include purchased viewership.
Depending on the platform, additional features can be added. In the case of Twitch, viewbotting can include automated chat features to lend credibility and the artificial viewers appear more organic. The inclusion of such features is generally platform specific.
Why Do People Viewbot on Twitch?
There are a few reasons that incentivize viewbotting, ranging from ego to money, although the latter seems a more compelling reason. Additionally, anyone can engage in Twitch viewbotting; it can happen without the streamer’s input or knowledge.
Let’s take a look at some of the reasons someone might viewbot:
- Discoverability: Twitch sorts streams by highest view count on the front page. Therefore, having the most live viewers allows a Twitch channel to be seen by more people. This is often how most people discover a new channel, as Twitch’s search feature leaves a lot to be desired.
- Favorable Positioning With Sponsors: The larger a streamer’s perceived audience, the stronger their negotiating position is with potential sponsors, allowing them to leverage more lucrative deals.
- Favorable Positioning Towards Peers: A larger audience can lend credibility to a streamer, potentially making them more desirable to collaborate with. This can help boost the number of organic viewers, making viewbotting less “necessary”.
- Delivering on Promises: Twitch organizations and “talent agencies” promise streamers that join them more lucrative sponsorships, deals, and exposure. As streamer xQc noted on Twitter that they also take a percentage of streamers’ ad revenue — by inflating the view counts, “they can sell massive ad packages in bulk and take in much higher sums.”
Viewbotting doesn’t have to be a permanent activity, either. It can be used to shore up a streamer’s view count until an organic community begins to develop, for example.
Furthermore, Twitch is top heavy, with a few creators that have exponentially more average concurrent viewers than smaller streamers, and with a lack of alternatives, smaller streamers work within the structure imposed upon them.
I would contend that Twitch’s front page, which prioritizes channels with the most live viewers, has unintentionally driven the appeal of viewbots. The platform also stipulates that a streamer must, among other things, “average of +75 viewers over a 30-day period” to be eligible for partnership.
How Does Viewbotting Work?
Viewbotting — or any automation in a similar vein — has a few hurdles to clear to be effective. The more strict a site’s anti-bot policies, the more hurdles there are. At its most basic level, a script has to be able to:
- Open a headless or visible browser instance
- Navigate to the stream page
- Handle login if needed (optional on Twitch)
- Keeping the page active
With no anti-bot measures in place, this is a very achievable set of goals with something like Selenium.
However, as more anti-bot measures are put in place, new challenges can appear. If the website uses browser fingerprinting, those have to be spoofed. If IPs are tracked, one has to use proxies or VPNs (with the former proving more effective at scale).
Account creation can be automated, although this is where most of a site’s anti-bot measures are concentrated with features like email verification, requiring a phone number, and other additional scrutiny.
This is to make no mention of the advantages that AI integration would bring, taking chat automation to a whole new level.
Crucially, though, automators have cleared all these hurdles and more in other settings. Sneaker and ticket sales are arguably some of the most hostile environments for automation, yet ticket scalping and sneakerbots are some of the most advanced examples of anti-detection technology.
The fundamental problem both the sneaker and ticket sale industry have to contend with is the following: The more aggressively you pursue bots, the more likely normal users are to be caught in the “crossfire”.
That fact isn’t lost on Twitch’s CEO, who in a statement on Twitter stressed that “we wanted to take our time to make sure we were not inadvertently filtering out real users.”
Viewbotting Isn’t Going Anywhere
There is an entire class of businesses whose main source of revenue is providing viewbotting services for every possible website and streaming platform — a very lucrative one, given the number of “viewers” who evaporated from August 21–23.
The industry will adapt, as other industries have, to the introduction of disruptive technologies or changes. The proxy industry has faced a few alleged crises, like the introduction of TCP/IP fingerprinting and the rollout of HTTP/3, and neither ended up being an existential threat as some feared.
In the immediate future, many companies offering viewbot services will be offline as they rush to make the necessary changes to their code. The number of bots did decrease, as we’ve seen. However, they’ll very quickly bounce back, as is evident from the promotional campaigns advertised on some viewbotting sites.
In the long term, Twitch may learn that cracking down more heavily on artificial viewership may be an exercise in futility. The demand for viewbots will not dissipate and providers will get better at hiding their product. As long as there are people willing to pay for views, there will be those with an incentive to inflate their numbers.
Conclusion
Only time will tell whether Twitch will continue to pursue its current course of action. It may find more success resolving the underlying issues fuelling the perceived necessity of viewbots is more effective, such as making it easier for smaller creators to be discovered.
Key Takeaways:
- Twitch’s crackdown on viewbots caused a short-lived drop in viewership on August 21–23.
- From discoverability to sponsorships, the incentives to viewbot remain unchanged.
- The more aggressive the anti-bot measures, the more likely real users will be affected.
- The viewbotting industry will (and likely already has) adapted to Twitch’s changes.
The drop in viewership was met with mixed reactions, some feeling vindicated in their suspicions while others were caught off guard by how much of the platform’s viewership was artificial. Unless the structural incentives change, viewbots will continue to remain a problem.
