From what I’ve seen, “deliverability” can be a tricky metric. You can always collect bounce data - that’s pretty standard. But anything beyond that (like adding internal tracking pixels or unique link IDs) often increases your chances of getting flagged as spam.
It’s a bit of a Heisenberg effect: You can’t measure it too closely without the measurement process changing the measurement. The more you instrument the emails, the less natural they look to spam filters, the worse your delivery rate.
I’d focus instead on the next layer of metrics: actual engagement - clicks and replies - since that’s where meaningful performance really shows up.
