Not OP, but use the tiered snapshots system extensively.
I personally just create datasets and then remove them from schedules if needed. Also, I have top-level datasets that segregate children datasets – something like Onset321, Backups or Slop. All these top levels have different active snapshot tiers. For example, you probably don’t need monthly snapshots for backups, because usually backup solutions have built-in versioning.
Realistically speaking, you only need to exclude the dataset from the longest tiers (like monthly or weekly), and it’s very unlikely that these tiers would make their snapshots right after the creation of the dataset.