Shale oil changed the supply side structurally, and the reason is not the volume so much as the response time. A conventional offshore project takes years from decision to first oil and then produces for decades at a fairly steady rate. A shale well takes weeks to drill and complete, declines steeply, and can be turned on or deferred within a quarter. That converted part of world supply from a fixed input into a price-responsive one.
The decline profile is the key mechanic. A shale well produces a large fraction of its total recoverable oil in its first year or two, then falls away sharply. Maintaining production from a shale basin therefore requires continuous drilling: stop, and output falls quickly on its own. That is the opposite of a conventional field, where stopping investment leaves production roughly flat for years.
This is why the rig count became a watched indicator, and also why it became a misleading one. Rigs drill wells; completions bring them online, and the two can diverge. A well drilled but not completed sits in the DUC backlog, and drawing that backlog down allows production to rise without any increase in drilling. Reading rig count alone as a proxy for future supply misses that entirely.
Productivity per well has also risen substantially through longer laterals and better completion design, which means a given rig count supports more production than it used to. Any historical comparison of rig count to output is therefore comparing across a moving technical baseline.