Hi @amvlad,
So the challenge with the Churn Risk prediction for use in segments and triggers is that often times businesses do not have a high percentage of repeat customers. Because of this, brand new customers to a business will likely end up as reflecting likely to churn once they make their first purchase. If we were to base things off of that information, churn risk could be misinterpreted or misunderstood leading to an email being sent out at a time where it potentially might not make sense to be reaching out to a customer i.e. a few days after they made their first purchase.
Instead of utilizing the churn risk prediction we recommend using Expected Date of Next Order instead of churn risk which should give a much better insight into creating a winback or re-engagement flow or segments. This would also take into account churn risk so Expected Date of Next Order could be used in place if this is important to your business.