The following is assuming the markerlight drone has a nearby drone controller. Markerlight drones without a drone controller are almost never cost effective.
I have found markerlight drones to be efficient if they shoot exactly once and then use saviour protocols. I'll explain this below.
Key assumptions:
#1: Tau's first two turns of shooting are the most pivotal. When evaluating markerlight models like pathfinders, I have assumed that a pathfinder will "pay for itself" after two turns. Thus per turn, a pathfinder's markerlight has a value of 4 points.
#2: A drone is more mobile than a pathfinder, but has roughly equal mobility to a sky ray. A stationary pathfinder markerlight costs 8 points. A mobile sky ray will hit a flying target on 3+. One sky ray markerlight is worth 10.67 points. Thus the drone's markerlight has a value of 8 points.
Implications:
If a ML drone shoots it's markerlight once, the markerlight will provide 4 points of value. If the drone then uses saviour protocols, it is the equivalent to only paying 6 points for saviour protocols.
A shield drone effectively costs 6.67 points per saviour protocol. Thus the markerlight drone is 11% more efficient at saviour protocols if it gets to shoot once.
Downsides:
#1. You must pay for a drone controller and thus sacrifice the option of using a different support system.
#2. If the drone dies before using it's markerlight, it is inefficient. You would have been better off using shield drones or pathfinders.
#3. If the makerlight drone dies after two (or more) turns of shooting, it is likely inefficient. This is because: It is often better to front-load your firepower to cripple your opponent asap. (I.e. It would be better to take 5 pathfinders than 4 markerlight drones if those drones were to survive 2+ turns.)
Conclusion:
Markerlight drones are quite efficient if you can use them correctly. You will need to fine-tune your army and pilot it with finesse, so that the markerlight drones die exactly when you want them to.