Retention is a great metric for products that cause frequent use (i.e., habit). So before you count, ask: with what natural frequency do people use/intend to use my product? Does retention really make sense for the value interaction in my product?
Retention is a lagging measure of success, so if you want to quickly see if there is value, then calculate engagement metrics (conversions to target action), which are proxy metrics for retention.
Retention is a metric that's best counted by segmentation to understand why it's there at all. Segmentation can be different: by channel, by geo, by JTBD, by type of customer, etc.
When you talk about retention, don't forget to add "at so-and-so month" and "going/not going to plateau".
Retention is good in comparison; it can rise or fall from cohort to cohort, and thus reflect the dynamics of your product’s success/failure.
Even if you have great retention, but the economics don't add up, the product doesn't have PMF.
Don't confuse User Retention and Revenue Retention; the latter can be greater than 100% (guess why).