That proportion is not too relevant for this graph. The graph is relative.

Even if a only a small percentage of Twitch users use Google search to access Twitch, the graph would be somewhat indicative of it's growth.

The question is: does the probability that a given user accesses Twitch over Google search go down as the user uses Twitch more often? Does it so more than on other sites, maybe because regular twitch use promotes other ways of engagement (phone app, console app, email notifications, twitter links from the streamers, etc)?

If 90% of new Twitch users reach Twitch over Google, 30% of one week old users and only 5% of one month old users, then a flat graph could actually indicate growth.

People who have just heard of twitch and want to know what it is? And whether those people eventually become twitch customers or not, their existence is a sign that twitch is still gaining mindshare in the broader community.

Hmm, this is interesting, for FB as a comparison:

https://www.google.com/trends/explore?date=all&q=facebook

Yep, corresponds almost perfectly with the rise of their mobile app.

Do you not know seemingly endless amounts of people who visit every site by googling the name?