Twitch Raids can cause a large influx of stream viewers. In this article, let’s do a deep dive on raids and use data to show the impact it has on Twitch streamers. Let’s start by defining exactly what we mean when we talk about raids.

A Twitch raid is an influencer directing their audience to another live channel at the end of their stream. The result is that the targeted channel has a massive influx of viewers. This is done as a way for influencers to promote other creators they like, as well as partake in what is a part of the Twitch meta culture.

The number of viewers can skyrocket with raids, boosting peak viewership. However, this boost is often temporary and does not result in an uptick in engagement or sustained viewership. A large number of viewers from this influx typically stop watching within minutes after the initial enthusiasm of being a part of a raid wears off.

We recently discussed how metrics like viewer-hours and average concurrent viewership (ACV) are robust against such activities. Using an example, let’s look at the problem of temporarily boosted viewership from Twitch raids and how well simple metrics like viewership fare against more complex ones like ACV.

Example: The LinusTechTips raid on PlantyTime

LinusTechTips had their live audience on Twitch and YouTube raid PlantyTime’s Twitch stream before ending their Weekly Analysis And News Show stream on March 27th, 2020
LinusTechTips had their live audience on Twitch and YouTube raid PlantyTime’s Twitch stream before ending their Weekly Analysis And News Show stream on March 27th, 2020

At the end of LinusTechTips' broadcast, the audience was directed to raid one of the host's friend, PlantyTime. PlantyTime was a relatively small channel prior to the raid, with peak viewership at 23 viewers. During the raid, PlantyTime's live stream viewership spiked to as many as 1,816 people.

PlantyTime’s viewership over the course of a Twitch raid, visualizing why peak viewership is non-representative of actual audience engagement/ retention.
PlantyTime’s viewership over the course of a Twitch raid, visualizing why peak viewership is non-representative of actual audience engagement/ retention.

Calculating the long-term impact of the LinusTechTips raid would be impossible with just peak viewership, whereas it is trivial with ACV. Using the below plot as reference, note the large deviation in viewership between PlantyTime’s streams after the raid. It is difficult to measure whether the raid was effective or not, or to infer anything about how the channel was affected long-term because of the sporadic viewership between streams.

However, measuring the raid’s effectiveness and long-term impact on the channel using ACV is far easier. The green line representing ACV for PlantyTime was steadily ~5 before the raid, and stabilized at ~60 after the raid. Even if a stream had a large number of peak viewers (such as during the subsequent raid in mid-April), because the influx of viewers did not engage with the stream, the ACV was largely unaffected.

PlantyTime’s daily peak-viewership vs. ACV
PlantyTime’s daily peak-viewership vs. ACV

ACV is robust against brief changes in viewership, because it is calculated as an average.

Beyond measuring viewership, ACV has other benefits too. It may be used with viewer-hours to find changes in how long people are watching streams for, or gauge a channel’s health relative to their growth. If you’d like to read more, we’ve talked about these in detail in our article discussing Twitch metrics.

Because of all of these factors, average concurrent viewership is one of the metrics we use at Gamesight to gauge how effective a streamer really is at engaging their audience.


About us

At Gamesight, we help PC and console marketers implement performance marketing techniques for their games. If you are seeking help setting up and measuring your campaigns, working with influencers, or would like to simply talk with us about this article, please reach out on our website!