How to track On-Time Delivery using SQL

How to implement on time delivery metrics in your business, and why they are key to customer satisfaction.

Archie Wood Jun 8, 2022 · 4 min read

Why are On-Time Delivery metrics so important?

When customers’ orders are not delivered on time, they may miss their delivery, or it may arrive too late to be useful.

Ensuring a high on-time delivery rate protects customer satisfaction, so it’s important to measure if customers are receiving their items on time.


A flawless handover

Definition

On-Time Delivery (OTD), sometimes called On-Time Performance, or On-Time Percentage (OTP) measures the percentage of orders that are delivered on time:

On\text{-} Time \ Delivery \ \% = {Number \ of \ On\text{-} Time \ Deliveries \over Total \ Deliveries} \times 100

Calculating On-Time Delivery

In order to calculate the OTD %, we need to know when the deliveries were meant to be delivered, and when they were actually delivered.

Assuming we are starting with the below data from 100,000 deliveries, which is stored in a database. For each order delivery_slot_start and delivery_slot_end mark the time period when the delivery was expected, and delivery_time is when the delivery was actually made.

select 
    order_id, 
    delivery_slot_start, 
    delivery_slot_end, 
    delivery_time 
from deliveries

Raw Deliveries

We first compare the delivery time to the start and end of the delivery slot to determine if the order was delivered on time, early, or late.

select 
    order_id, 
    delivery_slot_start, 
    delivery_slot_end, 
    delivery_time,
    case 
        when delivery_time > delivery_slot_end then 'Late' 
        when delivery_time < delivery_slot_start then 'Early' 
        else 'On-Time' end as delivery_status
 from deliveries

Deliveries Classification

From here, we can calculate the OTD % for each month in the dataset.

select 
    substr(delivery_slot_start, 0,7) as delivery_mon, 
    count(*) as total_deliveries,
    round(count(case when delivery_status = 'On-Time' then 1 end)*1.0/count(*),2) as on_time_pct,
    round(count(case when delivery_status = 'Early' then 1 end)*1.0/count(*),2) as early_pct,
    round(count(case when delivery_status = 'Late' then 1 end)*1.0/count(*),2) as late_pct
from ${deliveries_classification}

where delivery_mon not like '2022-01'
group by delivery_mon
order by delivery_mon

Monthly OTD

Presenting On-Time Delivery Metrics

On-time delivery metrics are typically presented as a time series, allowing the viewer to see the OTD % for each month.

We can use Evidence to plot a line chart to see how the OTD % changes over time:

Monthly OTD

We can also use an area chart to help show the proportion of deliveries that were late or early:

select 
    substr(delivery_slot_start, 0,7) as delivery_mon,
    delivery_status,
    count(*) as total_deliveries,
    round(count(*)*1.0/sum(count(*)) over (partition by substr(delivery_slot_start, 0,7)),3) as deliveries_pct 
from ${deliveries_classification}

where delivery_mon not like '2022-01'
group by delivery_mon, delivery_status
order by delivery_mon, delivery_status desc

Monthly OTD