It would be easier to answer this question if you mention the specific brand/model of device, as the answers will differ.
Zwave Devices: two different kinds of scenes
in most cases, I think these are probably Z wave devices using one of two completely different types of Z wave commands ( that’s why the brand/model matters).
The older method is just called “Z wave scenes”. In this case the “scene controller” (that’s a device class in zwave) issues a command directly to the target Z wave devices without telling the smartthings hub about it. An example of this type of device would be the old GE handheld remote. Typically you define the scene IDs by configuring the controller device.
The newer method is called “Z wave central scenes”. For these, the “central scene controller” sends a scene number to the SmartThings hub, and then you can have the hub issue other commands based on that seem number. The Homeseer switches which can do double tap and triple tap are using central scene commands. Pressing double tap on the top part of this which would be one scene number, pressing double tap on the bottom of this which would be a different scene number, etc. In many of these cases the central scene ID is established in the device’s firmware to apply to a specific device action.
Zigbee also uses “scenes” but in a somewhat different manner.
Devices that create scenes in their own native app
There are also some Wi-Fi devices that have scenes, like LIFX. (Hue scenes work in a similar fashion) In that case you set up the scene in the device’s native app first, and then have smartthings or IFTTT send a request to activate that scene.
Lutron devices also have scenes, but they aren’t currently supported in the official SmartThings/Lutron integration.
There are even some smartapps written specifically for smartthings like “scene machine” which have their own scenes and their own methods for setting them up.
so different devices, different answers
So the short answer is almost all home automation lighting devices have some kind of native “scene.” How you access it and how the scene IDs get assigned vary depending on the specific protocol and/or device. The most regularized is Z wave, which offers two different types of scenes, and there are a number of DTHs which will reference those.
So which devices were you looking at?