Key takeaways
- On-time delivery rate, first-attempt success rate, and failed delivery rate predict customer churn more reliably than total deliveries completed or average driver rating.
- Delivery KPIs should be segmented by driver and by day — a healthy overall average can hide one route or one driver dragging down customer experience.
- Businesses doing fewer than roughly 20–30 deliveries a day generally don't need a formal KPI dashboard yet; a shared spreadsheet and a weekly review are enough until volume forces the issue.
- A live dispatch platform gives you delivered-vs-failed counts and per-driver success rate automatically — cost per delivery and on-time-vs-promised-window need your own cost and promise data layered on top.
In this article
- Vanity KPIs vs. KPIs that actually matter
- The core delivery KPIs worth tracking
- How to actually calculate each KPI
- Where most delivery businesses go wrong tracking KPIs
- What Traksend surfaces automatically, and what you'll need to add
- Rolling out KPI tracking without disrupting operations
- When you don't need a formal KPI dashboard yet
Delivery KPIs are the small set of measurable indicators — on-time rate, first-attempt success, cost per delivery, failed delivery rate — that tell you whether your delivery operation is actually working, as opposed to just running. The ones that matter are the ones that move before a customer complains or cancels, not the ones that only look good in a monthly report. Most delivery businesses track too many numbers and act on too few of them.
Vanity KPIs vs. KPIs that actually matter
A vanity KPI is any number that goes up and makes the team feel good without telling you anything about customer experience or cost. Total deliveries completed is the classic vanity metric — it goes up every day you're in business, regardless of whether customers are happy. Contrast that with failed delivery rate or first-attempt success rate: both move in near-real time, both are directly caused by something your dispatch team can fix, and both correlate tightly with whether a customer orders again. The test is simple: if the number went up 10%, would you know what action to take? If yes, it's a KPI worth tracking. If you'd just shrug and say "great," it's a vanity metric.
The core delivery KPIs worth tracking
| Metric | Why it matters |
|---|---|
| On-Time Delivery Rate | Directly predicts repeat orders — customers rarely complain about a late delivery once, but a pattern of lateness is a leading driver of churn |
| First-Attempt Delivery Success Rate | A failed first attempt roughly doubles your cost on that order and delays the customer a full cycle — often the highest-leverage KPI to fix |
| Average Delivery Time / Time-to-Door | Reveals whether promised windows match reality — a growing gap between promised and actual time is an early churn signal |
| Failed Delivery Rate | Every failed delivery costs a re-attempt, a support conversation, and often a refund — the clearest line connecting operations to margin |
| Cost Per Delivery | Ties dispatch efficiency to profitability — rising cost per delivery with flat volume usually means routing or driver utilization has degraded |
| Customer Complaint Rate tied to delivery | Captures dissatisfaction that doesn't show up in on-time or failure data — a customer can receive an "on-time" delivery and still be unhappy about handling |
| Proof-of-Delivery Completion Rate | Protects you in payment and dispute disagreements — a drop is often the earliest warning sign of a driver cutting corners elsewhere |
How to actually calculate each KPI
On-Time Delivery Rate = (deliveries completed within the promised window ÷ total deliveries) × 100. Track this daily, not just monthly — a monthly average can hide a bad week that already cost you customers. First-Attempt Delivery Success Rate = (deliveries completed on the first attempt ÷ total delivery attempts) × 100 — a healthy range for most last-mile operations is 90–95%; below 85% usually points to address or time-window problems, not driver performance. Failed Delivery Rate = (deliveries not completed on the scheduled attempt ÷ total scheduled deliveries) × 100 — anything above 5–7% on a route or driver level is worth investigating immediately. Cost Per Delivery = total delivery operating cost (driver pay, fuel, vehicle, dispatch overhead) ÷ total completed deliveries over the same period — calculate this per route or per driver, not just company-wide, since a single inefficient route can quietly inflate your blended average.
Where most delivery businesses go wrong tracking KPIs
- Manual spreadsheet tracking — if a dispatcher has to manually log delivery times or attempt outcomes at the end of a shift, the data will be incomplete and eventually abandoned when volume picks up.
- Measuring averages instead of distributions — an average delivery time of 38 minutes sounds fine until you learn a quarter of deliveries are taking over an hour, and it's that slow quarter driving your complaints.
- Not segmenting by driver or route — a company-wide on-time rate of 91% can still mean one route is running at 60% while everyone else is near 100%, and the aggregate number will never tell you that.
- KPIs that aren't tied to a real workflow just become a report nobody acts on. If a dropping first-attempt success rate doesn't trigger an actual conversation with a specific driver, the metric is decorative.
What Traksend surfaces automatically, and what you'll need to add
Traksend's dispatcher dashboard tracks delivered and failed counts as orders move through the system, with per-driver success rate and a daily delivered/failed trend surfaced automatically, without a dispatcher having to log anything at the end of a shift — which gives you a live, always-current view of first-attempt success and failed delivery rate without manual counting. Proof of delivery — a confirmation code or timestamped photo — is captured at the point of drop-off, so a proof-of-delivery completion rate is something you can pull from that same data.
Two of the KPIs above take more than dispatch data alone. On-time rate needs a promised delivery window to compare against, which depends on what you quote customers at booking — Traksend gives you the actual delivery timestamps to measure against that window once you're tracking it. Cost per delivery needs your own driver pay, fuel, and overhead numbers layered on top of the delivery counts Traksend already gives you — it's not a cost accounting tool, so this calculation is one you'll run yourself using the volume and route data from the dashboard.
Rolling out KPI tracking without disrupting operations
Start by picking two or three KPIs — on-time delivery rate and first-attempt success rate are the highest-leverage starting point for most operations. Adding all seven at once tends to overwhelm a team that isn't used to reviewing metrics daily. Review the numbers with your dispatch team weekly for the first month, not daily — daily review at low volume produces noise, not signal.
When you don't need a formal KPI dashboard yet
If you're running fewer than roughly 20–30 deliveries a day, formal KPI tracking is probably premature. At that volume, a dispatcher or owner can usually hold the state of every delivery in their head, and a shared spreadsheet reviewed weekly will catch the same problems a dashboard would. The threshold isn't really about company size — it's about whether any single person can still notice a bad pattern without data.
Frequently asked questions
What is a good on-time delivery rate?
A realistic on-time delivery rate for most local courier and last-mile operations is between 85% and 95%. Rates consistently below 80% usually indicate a dispatch or routing problem rather than a driver performance issue.
How do you measure delivery performance?
Through a small set of KPIs — on-time delivery rate, first-attempt success rate, average delivery time, failed delivery rate, and cost per delivery — calculated from real-time tracking and proof-of-delivery data rather than manual end-of-shift logs.
Do small delivery businesses need to track KPIs?
Businesses running fewer than roughly 20–30 deliveries a day can typically rely on a shared spreadsheet and a weekly review instead of formal KPI tracking. Formal tracking becomes worthwhile once volume grows enough that no single person can track delivery quality from memory.
Does dispatch software calculate cost per delivery automatically?
Not on its own — cost per delivery needs your driver pay, fuel, and overhead figures, which a dispatch platform doesn't know. What it can give you automatically is the delivery volume and per-route data you divide those costs across.