Those offering a paid services should always ensure there is a benefit and convenience compared to piracy. Valve and Steam wouldn't be here if it was clumsy like in the beginning. Gamers like that getting a game is just a click away, that it installs and checks all the required dependencies, keeps the game file updated, allows you to preload and keep your game saves in the cloud (when possible).
At MIT, a few years after the arrest of Aaron Swartz, a prominent researcher from Mass General Hospital was threatened one afternoon with arrest by an MIT police officer for the alleged "crime" of using one of the library computer terminals for "longer than an hour."
It's worth remembering then that there are some out there who are using SciHub now because they don't want to go to jail for trying to access the same content through more "legally legit" channels.
This type of semi-automation would be very impractical and cost an order of magnitude more time if I had to dig through the probably non-existent API documentation individual publishers have, or try and scrape their I'm sure by now heavily obfuscated websites.
At the end of the day when considering the behavior of large numbers of heterogeneous users, UX is king for all data consumption. The pricepoint matters, but people would use SciHub even if traditional journals somehow moved to free publication... If the UX is better for SciHub.
It's also not like there was no piracy before scihub showed up. When I encountered a paper I didn't have access to back in the day, I would ping all of my friends at various universities / companies / hospitals and request that they check if they had access. If so, I would ask that they download it and send it to me. They would do the same when appropriate. Even doing this was more efficient than going through official channels, and every time this was done it represents another lost sale. From a user perspective, the piracy has nothing to do with money and is instead entirely about friction.