29 May 2026·7 min read

What is a good Amazon Cortex DCR score, and how do you improve it?

If you manage a UK Amazon DSP, your Amazon Cortex DCR score is the metric that can define your entire week. A single bad day — a wave that ran over time, a batch of returns, a driver who went off-route — shows up on Monday as a percentage drop that feels impossible to trace. This guide explains exactly what DCR measures, what separates a Fantastic score from a Fair one, the most common reasons it falls, and the specific steps you can take to push it back up before the next weekly scorecard lands.

What DCR actually measures — and why Amazon weights it so heavily

DCR stands for Delivery Completion Rate. It measures the share of your assigned stops that were actually delivered successfully during the wave. If your driver was allocated 120 stops and completed 117, their DCR for that route is 97.5%.

Amazon weights DCR heavily because it is the most direct proxy for customer promise fulfilment. Every stop that does not complete either generates a redelivery attempt (which costs Amazon money), a customer complaint (which increases your DNR exposure), or a returned parcel (which the customer must then chase). Across a fleet of ten drivers doing six days a week, even a 1% DCR miss compounds quickly.

DCR is calculated weekly across your entire DA fleet. It does not average out neatly — a single driver completing three short waves with several missed stops can pull down a fleet that is otherwise performing well. That is why understanding DCR at the individual driver and route level matters far more than watching the fleet headline figure alone.

What a good Amazon Cortex DCR score looks like

Amazon's Cortex scoring bands work roughly as follows (exact thresholds vary by programme and region, but UK brackets align broadly with these figures):

Fantastic Plus: DCR at or above 98.0%
Fantastic: DCR between 96.5% and 97.9%
Great: DCR between 95.0% and 96.4%
Fair: DCR between 90.0% and 94.9%
Poor: DCR below 90.0%

Most DSPs aiming for Fantastic Plus need every driver consistently above 97%. That sounds achievable until you factor in long days, Same-Day Delivery waves, bad weather, and the occasional driver who ends a shift early.

The target to aim for is not 96% — it is 98% and above, because that is the buffer that absorbs your worst-performing route on your worst day of the week and still keeps you in the top band. Anything below 95% starts to attract attention from your Station Operations Manager and can affect your zone and route allocation in subsequent weeks.

The five most common reasons DCR drops

Understanding why DCR falls is the first step to fixing it. These are the five patterns seen most often across UK DSPs:

Waves that run over time. When a driver has 130 stops and a 10-hour shift, the maths do not work. Stops that cannot be completed before the wave closes automatically count against DCR. Over-loading happens when route planning does not account for realistic completion rates, traffic density, or stop complexity.
Returns to station with packages. If a driver ends the day with undelivered parcels rather than attempting every stop, those count directly against DCR. This is often a discipline issue, but it can also be a training one — drivers who do not understand the consequence of leaving stops undelivered will repeat the behaviour.
Failed attempts at problem addresses. Some addresses generate repeated failures — gated properties, businesses that close early, access codes that do not work. Each failed attempt on an address that was never going to be deliverable drags the score without the driver doing anything wrong.
Same-Day Delivery (SDD) wave compression. SDD waves have tighter time windows and less tolerance for exceptions. A driver handling both standard and SDD on the same day is more likely to leave stops incomplete because of the competing time pressure.
Inexperienced drivers on dense urban routes. New drivers in city centres take longer to locate addresses, struggle with access, and are more likely to miss difficult deliveries. Assigning newer drivers to routes with a high stop density and tight timing is a reliable way to pull DCR down.

A per-driver playbook to fix it this week

DCR improvement happens at the driver level, not the fleet level. Here is the sequence that works:

Pull the last four weeks of DCR data for every driver individually. Sort it lowest to highest. The bottom three or four drivers will almost always account for more than half of your DCR gap. That is where to focus first.

For each driver below 96%, work through these steps:

Identify the specific routes or waves where they missed. Was it the same route each time? If so, the route is likely over-stopped, not the driver.
Check whether incomplete stops fell at the start or end of the wave. Late-day misses suggest a time management issue. Early misses point to address-level problems.
Reroute or rebalance if the route itself is the problem. Some routes genuinely cannot be completed in the time window allocated. No amount of driver coaching fixes a structural mismatch.
Run a one-to-one conversation — not a group briefing. Drivers respond differently when they see their own score rather than the fleet average. A targeted conversation about their specific numbers lands differently from a Monday morning announcement about DCR being down.

Track progress daily through the week. By Thursday you will know whether the changes are working or whether a different intervention is needed before Monday's scorecard.

How to spot a falling DCR before the weekly scorecard lands

The weekly scorecard is a lagging indicator. By Monday, the damage is already done. DSPs that consistently hit Fantastic Plus have real-time or daily visibility into the signals that predict DCR — not the score itself, but the events that produce it.

Indicators to watch mid-week:

Waves closing with packages still on the van. If your OSM can see at 4 pm that a driver has 15 stops remaining with 45 minutes of wave time left, intervention is still possible — a phone call, a route swap, a time extension request to the station.
Return-to-station (RTS) events. Every RTS event with packages still on board is a direct DCR miss. Each one should generate an alert the same day so it can be investigated and addressed before it happens again the next morning.
Incomplete-stop patterns clustering around specific drivers. One driver missing five stops across three days is a different problem from three drivers missing two stops each. The former needs an individual intervention; the latter probably needs a route or timing review.

Linking your Cortex data to a live driver performance tracking tool is what separates reactive Monday reviews from proactive Wednesday fixes. The DSPs that hold Fantastic Plus week after week are not lucky — they see the problem on Wednesday and resolve it before the wave closes.

Frequently asked questions

What is a good DCR score on Amazon Cortex?

A DCR of 98% or above puts you in the Fantastic Plus band, which is the target for DSPs who want maximum flexibility on route awards and zone allocation. Fantastic (96.5%–97.9%) is acceptable but leaves no buffer for a difficult week. Anything below 95% (Great band or lower) starts to attract scrutiny from your Station Operations Manager and can affect future route and zone assignments.

Why is my DCR low even when my drivers are completing their routes?

DCR misses do not always mean drivers are skipping stops deliberately. Common hidden causes include waves timed too tightly so late stops close before they can be completed, Same-Day Delivery schedule pressure bleeding into standard routes, and problem addresses that generate automatic fails regardless of how many attempts are made. Check your incomplete stops at the route level rather than the fleet level to find the pattern — it is nearly always concentrated on a small number of routes or drivers.

How often does DCR update in Amazon Cortex?

Amazon Cortex updates its DA scorecard metrics weekly. The scorecard visible on Monday reflects activity from the previous week (typically Monday to Sunday). There is no live DCR figure within Cortex itself, which is why DSPs who need intra-week visibility have to pull metrics via a connected tool rather than waiting for the weekly reset. By the time you see a low score in Cortex, you are already a week behind.

Does one driver's low DCR affect my whole fleet score?

Yes. DCR is calculated across your entire DA fleet. A single driver completing three partial waves in a week can pull your fleet DCR down by 0.5–1.0 percentage points depending on fleet size. This is why per-driver visibility matters more than the headline fleet average — the fleet number hides the individual contributors.

Can I dispute a DCR miss if the stop was undeliverable?

In some cases, yes. Amazon allows exception requests for certain stop types — gated properties, incorrect addresses supplied by the customer, and customer cancellations. The process varies by station, but your Station Operations Manager is the first contact. Document the specific stop details and the reason for non-completion at the time it happens, not after the scorecard has landed — retrospective disputes are harder to support without same-day evidence.

If you want daily DCR visibility without rebuilding your tracking in spreadsheets, DSPOps connects to Amazon Cortex and surfaces every driver's completion rate mid-wave — not after Monday's scorecard. It is Amazon DSP management software built specifically for UK operators, with a 14-day free trial and no credit card required.