Bangladesh is one of the fastest growing e-commerce markets in South Asia. Mobile internet penetration is rising. Digital payment infrastructure is expanding. New buyers are coming online every month. And yet a quiet operational problem sits underneath all of that growth that does not get talked about enough.

Failed delivery.

In Bangladesh e-commerce, a package that leaves the warehouse does not always reach the buyer. The buyer either refuses it at the door, is not available when the rider arrives, or never intended to accept it in the first place. The platform absorbs the full logistics cost either way. The seller loses the sale. The rider made a wasted trip. And the revenue that the GMV number promised never materialises.

This is not a small problem. And in Bangladesh specifically, it is harder to solve than most markets acknowledge.

Why Bangladesh is structurally different

Most frameworks for managing failed delivery risk were built with a different market in mind. They assume standardised postal addresses, mature digital payment habits, and a buyer base with experience placing and cancelling online orders. Bangladesh has none of these in full yet.

Street addressing in Bangladesh is inconsistent outside of major urban centres. A rider navigating a delivery in Mirpur or Mohammadpur is working with address information that often lacks a precise building reference. That is not a buyer behavior problem. It is an infrastructure problem that no risk rule can solve.

Cash on delivery accounts for a disproportionately high share of orders in Bangladesh compared to even comparable markets in the region. That means the financial exposure per order is higher here. Every failed delivery in a COD-heavy market is not just a logistics cost. It is a full round trip cost with zero revenue recovery. When this happens at scale the impact on gross-to-net margin is significant.

The buyer composition challenge

Bangladesh e-commerce is acquiring new buyers faster than it is educating them. That gap creates a specific type of failed delivery that is easy to misread as fraud but is actually something else entirely.

A first time buyer who placed a COD order and changed their mind does not necessarily know how to cancel it. The cancellation flow requires navigating an app interface that is still unfamiliar to a large portion of new users. The path of least resistance is not cancelling. It is saying no to the rider when they arrive.

This buyer is not a bad actor. They are an under-onboarded buyer. Treating them the same way as a deliberate repeat rejecter is one of the most common and costly mistakes in how platforms manage failed delivery risk in Bangladesh.

Where failures actually concentrate

When you look at the data on when failed deliveries happen relative to when a package enters the logistics pipeline, a clear pattern emerges. Successful deliveries concentrate in the first two to three days after the package is picked up for transit. Failed deliveries concentrate disproportionately in packages that have been in the pipeline for more than a week.

This matters because it means a significant share of failed delivery cost in Bangladesh is not driven by buyer intent at the point of checkout. It is driven by packages aging in the last mile without a successful delivery attempt being made. The buyer who was willing to accept the package on day two may not be available or willing on day nine.

That is a logistics timing problem. And a risk rule that blocks a buyer because of failures driven by logistics timing is solving the wrong problem.

What actually needs to change

Managing failed delivery in Bangladesh requires separating three populations that most platforms currently treat as one.

The first is buyers who are genuinely high risk. They have a pattern of deliberate rejection, sometimes across multiple accounts on the same device. These buyers need a hard intervention.

The second is new buyers who failed because they did not understand the process. They need education and a proactive contact before their next delivery, not a restriction.

The third is buyers whose packages failed because they aged in the logistics pipeline before a delivery attempt was made. They need a logistics fix, not a risk action.

Bangladesh e-commerce will not resolve its failed delivery problem by writing more aggressive risk rules. It will resolve it by understanding which population it is actually dealing with in each case and designing the right response for each one.

That distinction is where the real work is. And in my experience it is the distinction that most platforms in this market have not yet made precisely.