The Delivery economy and its new building blocks — Part II

The last post outlined challenges doing last mile for a growing demand of doorstep-commerce. Managing delivery operations is messy at scale and becomes expensive quickly due to poor customer access, low biker productivity & high cost of last mile transactions. Whether it is a waste-management company doing pickups, a healthcare/online-pharmacy mobilising health services or medicines at home or routine ecommerce deliveries; last mile needs a smart delivery management apparatus.

Deliveries are set to balloon to 40bn globally by 2025 with charges dropping below a dollar. Google express, ubercargo, amazon, instacart and some bigger stores like Target have already begun offering delivery services on subscription with a delivery fee of 80 cents or less. This means delivery costs would have to reduce 4 times from current levels or productivity have to go up 150% or both. Keeping this future in mind we are building an operations stack only for last mile that combines the agility to plan between an on-demand, same-day or next-day delivery with the AI that can boost delivery output and biker productivity.

The stack, that is built like a SaaS on cloud, abstracts out the entire delivery planning practice as a separate function focusing on the following three things for the desired outcome :

Automated Workflows: Our SDKs, built for all types of server and db architecture, links the db that stores delivery info to our stack within minutes. Built with a novel webhook (auto-call back) these SDKs keep pushing any new delivery entry to the stack as soon as they show in client’s db. Post extraction, automated workflows connect delivery tasks like planning, scheduling, sorting, dispatch and tracking of shipments, that are mostly done manually to a single order lifecycle. Both can potentially reduce delivery planning time by 35% freeing up more time from non-essential tasks and help run last mile in a lean & mean manner.

Smarter Location Mapping: NLP-enabled address contextualisation could be harder to spell, simply put our own version of a geo-mapping can map a raw and misspelt customer address to its physical location with higher accuracy. This geo-mapping service learns from the common spelling, landmarking and locality contextualisation habits user employ writing an address, only to arrive at more accurate delivery end-locations that helps shipments reach to users faster.

Delivery Planning & Routing: Our routing engine helps company A/B test delivery schedules either improving biker productivity or reducing capacity requirement for a fixed throughput. In easy terms, a 500 delivery load for a day could either have fewer than available bikers complete all deliveries or have biker man-hours saved creating additional capacity. The AI can suggest planners to pick a delivery mode depending on load volume & distribution and in both cases reduces/creates capacity by 25–30%.

We are supporting an interesting clutch of Ecomm specialists, drug delivery, healthcare services, mobile repair & waste-management companies whose last mile operations have grown and became complex with increasing demand. The end goal is that with the 4bn-5bn deliveries that happen alone in India by 2025 annually, AI and ML could possibly a big role in understanding how high customer satisfaction can go hand in hand in scaling delivery operations so that costs do not spiral out of control

Like what you read? Give Transporter a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.