Mark,

Imagine you had an illustration of a stove/cooker.

You make the base of the stove first - with burners and the racks for putting cookware on.

Next you add a kettle to put on top at the back.

Now you add a cooking pot at the front.

It looks wonderful.

Now your client says it would look better with the kettle at the front and pot at the back. You have to rejig the stacking order of every one of those layers (if you didn't group them).

Then the client says that they want images of the stove with just a kettle, just a post and just the stove. And now they want two pots, one to be red.

Now you can do all this by careful grouping, but all the while you are in danger of messing things up because all these parts sit one on top of the other and if you want to add a pot, you can accidentally move the kettle or change it.

When you have layers you can distribute all of these parts to different layers - one for the stove, one for the kettle, one for each pot.

Now you can control each object through the layers. You can draw the stove and lock that layer. Then you can draw the kettle on the stove, but the stove is protected. You put the kettle on one layer, lock it then draw the pot on a new layer.

If you want a new pot, duplicate the layer.
If your kettle needs to move in front of the pot, change the layer order.

You can rejig the artwork on an isolated layer, knowing you won't screw up your other work.

I don't know how people can make complex images and avoid layers without making a lot of work for themselves.