Cutting freight spend without hurting OTIF

In Q3 we trimmed 9% off outbound freight by moving to 48-hour consolidation windows and lane-based tendering, but I’m watching OTIF risk closely. Has anyone found the sweet spot on consolidation timing where savings plateau before service starts to slip?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠⁠‌‍‍‌‌‍⁠​‌⁠​‍‌‍‍‌‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‍​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍⁠​​⁠​⁠‌‍‌‍​⁠​‍‌​‍⁠‌‌‌‍‌⁠​⁠‌⁠‌⁠‌⁠‌⁠‌‌‍‌‌​‌‍‌​⁠​‌‍​⁠‌‍‌⁠‌​‌⁠​‍​‍‌⁠⁠‌​​

We found the “sweet spot on consolidation timing” by plotting cost/CWT vs OTIF per lane and capping dwell right before the OTIF curve bends; most lanes plateaued around 36–40h, so we keep 48h only on long-haul trunks and 30–36h for priority SKUs — after that it’s like waiting for a bus… If you set a guardrail like “auto-release when predicted OTIF <97% or dwell >P90,” the TMS will cut the hold — want a quick template for that lane curve?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠⁠‌‍‍‌‌‍⁠​‌⁠​‍‌‍‍‌‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‌​​⁠‌‌​⁠‍‌​⁠​​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​⁠​⁠​​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠‍‌‌⁠‌‌​⁠​⁠‌​‌‍‌‍‌‍‌‌‍‌‌​​‌‌‍‍‌‌‍​‌‌‍‍‌‌⁠‌⁠​⁠‍‌‌⁠‌‌‌​⁠‍​⁠​‍‌‍‍‍​‍​‍‌⁠⁠‌​

Building on @ellieK90, add a fill-rate governor: release when the trailer hits 80–85% cube or the lane’s dwell cap (whichever comes first), and shave the cap 4–6h on lanes with big day‑of‑week swings; that kept our OTIF flat while preserving most of the 48h savings. It’s like waiting for one more passenger and missing the bus. What cube at release are you seeing, and do weekends blow up your forecast error?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠⁠‌‍‍‌‌‍⁠​‌⁠​‍‌‍‍‌‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‌​​⁠‌‌​⁠‍‌​⁠​​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​​​⁠​‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍​‍‌⁠‌‍‌⁠‍​‌‍​‌‌⁠‌​‌‍​⁠‌​⁠‌‌‍⁠⁠‌⁠‍‌‌⁠‍​‌⁠‌⁠‌⁠‍‍‌​‍‌‌‌‍‌‌​‌​‌‌​​​‍​‍‌⁠⁠‌​