The NDR Breakdown: What Your Courier's Reports Are Actually Telling You

Your weekly NDR report from Shadowfax or Delhivery looks like data. It's actually a story your courier wants you to believe — and the gap between the story and reality is where most ops leads stop being able to manage RTO.

If you've spent any time looking at NDR (Non-Delivery Report) data from an Indian courier, the cause distribution is roughly this:

Customer refusal/rejection: ~37% of total NDRs
Customer unavailable: ~28% of total NDRs
Invalid/unreachable contact number: ~12%
Out-of-delivery area (ODA): ~2%
Other / unattempted: remainder

The "customer unavailable" problem

That 28% "customer unavailable" line is where things go sideways. Here's what's actually happening behind it:

Delivery agents in India are overloaded. Route optimisation is poor. Distribution hubs run on space pressure — incoming parcels need somewhere to go, so unattempted ones get marked NDR to clear the bay. The agent's incentive structure rewards number of attempts logged, not successful deliveries. So the path of least resistance, for an agent running behind, is to mark the parcel "customer unavailable" without leaving the hub.

Industry estimates put this at 12–15% of all unsuccessful deliveries. Post-delivery surveys are even more damning — when brands send "did you receive your parcel?" surveys to addresses already marked "delivered", up to 70% respond no.

Why "customer unavailable" specifically

It's the cleanest cover. "Customer refused" implies you talked to a customer. "Invalid number" can be cross-checked. "ODA" gets flagged for re-zoning. But "customer unavailable" requires no evidence — and it's not the agent's fault, it's the customer's. The category is structurally the place that fake attempts go to hide.

How to interrogate your own data

Three things you can do this week with no new tools:

1. Compare NDR rates across couriers in the same pin codes. If Delhivery's "customer unavailable" rate is 22% and Shadowfax's is 38% in the same zone, that's not customers — that's couriers.

2. Cross-reference with your repeat-buyer data. If a pin code has high "customer unavailable" rates AND high repeat-buyer percentage, the customers in that zone are clearly available. The courier isn't.

3. Run a 30-day post-delivery survey. WhatsApp, simple yes/no. Compare against your courier's NDR data. Any gap is your fake-attempt rate. Detailed walkthrough here ↗.

What to do once you know

The conversation with your courier changes immediately. You're no longer asking them to "improve service" — you're showing them a 30% gap between their NDR data and your survey responses, by zone, by week. That gets attention.

The harder fix is structural: fake attempts are a feature of the gig-economy last-mile model. The only durable solution is removing the failure mode entirely — either by switching to own-rider hyperlocal dispatch or by routing to a pickup locker that doesn't require the customer to be home. Both of which NanoHub does.