The challenge
Port authority leadership faced a strategic question before committing to a planned 40% capacity expansion: would adding physical capacity actually achieve their throughput targets, or was the underlying problem operational? The port was handling 9.2 million TEUs annually with average container dwell time of 6.8 days — 2.3 days above target and well above global benchmarks. The congestion was systemic: vessel arrival data was unreliable, berth allocation relied on manual spreadsheets, container tracking used 23-hour-lag batch data, and six operational departments worked without a shared operating picture. Leadership needed an operational intelligence programme that would address the efficiency gap and provide the foundation for physical expansion.
Discovery
A 10-week operational assessment revealed that the port's congestion was caused by information asymmetry, not physical capacity. Vessel arrival data was unreliable, container tracking arrived 23 hours late, and six departments had no shared picture of port state. Our diagnostic concluded that adding berths without addressing operational intelligence would increase the congestion problem, not resolve it.
Our approach
Diagnostic and problem definition: conducted a 10-week operational assessment mapping all data flows, decision processes, and bottlenecks across berth management, yard operations, gate management, and customs coordination
Platform architecture and integration design: designed the unified port intelligence architecture, integrating the legacy terminal operating system, berth planning system, customs platform, and carrier EDI feeds through a real-time data layer
Vessel prediction programme: built an arrival prediction model processing AIS data, historical carrier schedule adherence, weather data, and canal transit information to deliver 72-hour probabilistic arrival windows
Berth optimisation implementation: designed and built berth allocation optimisation, replacing manual scheduling with a system that optimises berth assignment, crane scheduling, and yard pre-positioning based on predicted arrivals and cargo profiles
Dwell management programme: developed a container dwell time prediction model identifying at-risk containers 48 hours before predicted breach, triggering proactive notifications to shipping agents and customs
Executive intelligence layer: designed and built a unified port operations dashboard providing real-time throughput, congestion risk, and performance data across all operational domains
The Outcome
The engagement delivered a real-time operational intelligence layer connecting all port systems into a single view of port state — from vessels in transit to containers in the yard and trucks at the gate. Berth planning shifted from reactive scheduling to predictive optimisation, enabling the port to prepare resources before vessel arrival rather than after. The dwell management programme addressed the administrative delays that were the primary driver of extended dwell times. Together, these capabilities reduced average dwell from 6.8 days to 3.1 days — surpassing the port authority's 3.5-day target — and enabled a 37% throughput capacity increase without physical expansion.
The results
“We knew we had an operational efficiency problem — but we did not have a clear picture of where the losses were coming from. CityPulse gave us that picture in real time, and then the tools to act on it. The dwell time results went beyond what we committed to our board.”
