I remember how frustrating scanning was. Then I had a moment of light.
Firstly, I assume you’ve seen: https://wiki.eveuniversity.org/Probing_In_Simple_Steps
The way scanning works is each probe gets a directional bearing to the location of the anomalies. Now that’s a little errored based on your skills and the resolution of the probes. The location of the anomaly of interest is thus localised by where the bearings cross. The further from the probe the less accurate, the greater range the probe is working at the noisier it is and thus less accurate so you are progressively closing the target down.
So what you are trying to do is position the probes around the “rough area of the anomaly” but in such a way that the bearings are all nice and wide and you get a clean set of directions rather than having all the probes bunched. It’s like trying to use a compass and map to work out a location - you want bearings a good angle apart, ideally 90 degrees, rather than all very closely together when inaccuracies start to build up. But backwards (your probes are the landmarks taking bearings to the target and drawing them on the map - the target is where they cross).
It’s tempting to think it’s just about raw power of the probes and heavily overlapping the spheres (I did!!!), but this puts the probes in almost the same place and they all just report the target on almost the same bearing and you have no idea where it actually is. Swing some out to the side and you get a cross-bearing to the target. This is passive radio direction finding not Radar.
Drawing - imagine the antenna are probes looking for the anomaly that’s giving of a bit of signal. Imagine the uncertainty if the direction lines are fuzzy.
OK, explained really badly there - basically, don’t close the probes up so the spheres all overlap, move the probes around a bit so they get different angles on the likely target and you’ll hopefully have more luck.