How do you measure a freight forwarder's service quality?

Most freight tenders compare prices and promises. Almost none measure service quality, because there is no shared set of metrics: every forwarder presents the figures that suit it, with its own definition. Here are the six metrics that make a real comparison possible, and the questions that tell you whether a claimed figure means anything.

The six metrics that matter

1. On-time against the quoted transit time. The share of shipments delivered within the transit time sold in the accepted quotation. This is the central metric: “reliable” means “hits its dates”. It is computed shipment by shipment, comparing what was sold to what happened — never by comparing two averages.

2. Incident rate, broken down by type. The share of shipments that hit at least one disruptive event, with the breakdown: rollover or cargo split, booking to redo, pre- or on-carriage to rearrange, customs inspection, document to correct, cargo lost or damaged, invoicing dispute. The breakdown tells you more than the headline rate — it shows where the provider is weak.

3. Responsiveness. The share of client messages that got a first human reply within a stated threshold, during business hours. Pin down the threshold and the time zone: “within 30 minutes during the handling office’s business hours” is not the same claim as “within 4 hours”.

4. The gap between quoted and invoiced price. The share of shipments invoiced within a tight band around the quote. Some gaps are unavoidable and legitimate: demurrage and detention, storage, truck waiting time, customs inspection. They must be excluded from the calculation and stated as such. The rest — rate error, forgotten line, stale rate card — is a measurable failure.

5. Per-shipment satisfaction. The share of shipments rated positively by the client at delivery. Simple, but only if the response rate is published alongside: 98% satisfaction on 3% of shipments rated means nothing.

6. Net Promoter Score. Promoters minus detractors. Useful for the trend, fragile when few people answer: always ask for the response rate.

Checking a claimed figure: four questions

A percentage on its own is never verifiable. Four questions are enough to tell a measured figure from a manufactured one.

• What is the denominator? 96% of what: of all shipments, or only of those the client bothered to rate? And what share of the activity does the measurement actually cover? A provider can give you that coverage rate without disclosing its volumes, which are commercial data.
• What is the scope? Which entity, which corridors, which modes. A group figure presented as a local one is a false figure.
• What is the period? A year-to-date number smooths what a quarter reveals. Ask for both.
• What is excluded? This is the most revealing question. An honest methodology writes its exclusions before publishing its result: cancelled files, client-caused delays, shipments in transit, cost lines that cannot be anticipated.

A provider who answers those four without preparation genuinely measures. A provider who offers to come back to you does not.

Writing them into a contract

Three rules make a service quality commitment usable:

• Write the definition, not just the target. “95% on time” commits to nothing without the formula, the denominator and the exclusions.
• Set a quarterly review. A month is too noisy; a year is too late to correct anything.
• Provide for access to anonymised raw data. A metric you cannot recompute is a statement, not a measurement.

Why almost nobody publishes these figures

Because publishing them means agreeing to show the bad quarters too. Most forwarders prefer hand-picked references and satisfaction rates with no denominator. That is exactly what makes the exercise discriminating: a provider willing to define its metrics, measure them across its whole scope and publish them at a regular interval is taking a risk the others decline.

At OVRSEA these metrics are tracked internally at group level, with their definitions, known biases and exclusions documented. A quarterly publication is in preparation, on the principle described here: the figure, its formula, its scope, its coverage and its extraction date. Volumes are not published: they are commercial data, and the coverage rate is enough to make a percentage interpretable.