I'd do something like...
Overall Wild Surge Effect (3d6):
3-6: No, and...
7-8: No.
9-10: No, but...
11-12: Yes, but...
13-14: Yes.
15-18: Yes, and...
No, and...
The spell works backwards if that would hurt you / your allies. Otherwise, the spell fails and something else bad happens to you / your allies, and the bad thing is a spell of an equal level to the spell you cast.
No.
The spell fails in a way that wastes your action but doesn't particularly hurt you / your allies.
No, but...
The spell fails, but something mildly beneficial happens (to you / your allies). The effect is a spell two levels lower than the spell you cast.
Yes, but...
The spell works, but also something bad happens (to you / your allies). The effect is a spell two levels lower than the spell you cast.
Yes.
The spell works, but something about it is random: perhaps the energy type changes, or it has a different rider condition, or the duration is randomly determined, or the material produced is unusual. Think of this as a version of the spell that could be researched (and would be balanced for the level).
Yes, and...
The spell works, and something about it is better than normal (by a random amount): for example, maybe it inflicts more damage than usual (of a random type), or maybe it summons an few small Elementals of the energy type to serve you for 1 minute.