The standard safety stock formula (Z × σLT × D) works well when your lead time has a predictable standard deviation. In practice, especially with suppliers shipping from Asia or Europe through Gulf ports, lead time variance can be extreme — a shipment that normally takes 28 days can take 45 days due to port congestion, vessel changes, or customs holds.
How we've adapted:
We use a modified formula that replaces standard deviation with the difference between maximum observed lead time and average lead time over the past 12 months. It's more conservative but has prevented stockouts on our fastest-moving SKUs.
Safety Stock = (Max Lead Time − Average Lead Time) × Average Daily Sales
We then apply a service level multiplier (1.65 for 95% service level) on top.
The result is higher inventory carrying cost but we've reduced emergency air freight by about 40% since switching to this approach.
The tradeoff is working capital. For expensive SKUs this formula ties up significant cash. We apply a lower multiplier for high-cost, slow-moving items and accept a lower service level on those.
What formula or approach are you using? Has anyone found a good middle ground between cash conservation and service level?

