Measurement & Reporting · 6 min read · September 25, 2026 · Hafsa Tariq, gaash.ai
Operating figures a buyer can check, each with its period and definition
Publish each operating figure as your own measurement, with its definition, its period and the date it was measured: cut-off times per service, receiving and dispatch windows, throughput or capacity, on-time performance. Say how each one is computed, what it excludes and when it will next be updated. A figure without a period and a definition is a claim, and an answer that repeats it repeats a claim.
A figure without a period and a definition is a claim
Compare two lines. "[on-time rate] on-time delivery" gives a buyer nothing to check: on time against which date, for which shipments, over which months, measured when? "[on-time rate] of [n] deliveries between [date] and [date] arrived by the committed date; excludes [n] shipments the client rescheduled; measured [date]" can be checked, compared with a client's own records and quoted without losing its meaning.
The second line is also the one worth quoting. On the benchmark in Aggarwal et al. (KDD 2024), sources that added statistics, quotations and citations gained up to about 40% in visibility within generated answers. The paper measured sources in general, not logistics providers, so read it as the case for publishing numbers with their conditions attached, which is what gives a passage its citability. The article on publishing your own numbers covers the general case; this article covers the four a logistics buyer asks about.
Cut-off times, per service and per site
A cut-off is a promise with conditions, and each condition belongs in the sentence: the time, the time zone, the event it refers to (an order in your system, a truck at the dock, documents submitted), the service, the site, and what happens after it passes. Receiving and dispatch windows are the same kind of statement for the dock.
| Figure | State it like this | Not like this |
|---|---|---|
| Order cut-off | "Orders received in our system by [time] [time zone] on a business day ship the same day from [site]" | "Same-day shipping" |
| Cargo and document cut-off | "Cargo at [site] by [time] on [day], documents by [time] on [day], for the [day] departure to [destination]" | "Weekly departures" |
| Receiving window | "Inbound receiving [days], [time] to [time] [time zone], by appointment booked [n] hours ahead" | "Round-the-clock operation", with no receiving hours |
| Dispatch window | "Outbound trailers leave [site] between [time] and [time]" | No time at all |
| Holidays | "Cut-offs for [holiday period] posted on this page by [date]" | Changes announced only by email |
Throughput and capacity: say what is counted
Throughput figures break down on the unit. "[n] orders a day" means one thing if an order is one line and another if it is forty. Name the unit (orders, order lines, units, pallets), the statistic (the median day, the busiest day, an average over [period]), the site and the period. Capacity is the ceiling and utilization is how much of it is in use: publish the first, and the second only if you will keep it current.
- "[n] order lines picked per day at [site], median over [period], measured [date]"
- "[n] orders shipped on the busiest day of [period]"
- "[n] pallet positions, [n] of them at [range] °C, counted [month year]"
- "[n] dock doors and [n] trailers in service on [date]"
On-time performance: the definition sets the number
An on-time rate is the result of six choices, and two providers with the same operation can publish different rates by choosing differently. That is why a buyer asks for the definition first. Publishing it beside the figure answers the question before it is asked.
| Choice | Options | What to publish |
|---|---|---|
| The date it is measured against | The date you committed to, the date the client asked for, an estimate | Which one, in the same sentence as the figure |
| The event that counts | Dispatch, delivery, proof of delivery | Which event, and where its timestamp comes from |
| The tolerance | The same day, within [n] hours, inside a delivery window | The tolerance, stated once and kept |
| The unit | Shipments, orders, order lines | The unit, and the count behind the rate: [n] of [n] |
| The exclusions | Client reschedules, refused deliveries, [other] | Each exclusion, and how many shipments it removed |
| The period | A month, a quarter, a year | The first and last dates, and the date it was measured |
Write the definition in the words your contracts use. A published rate that uses the same definition as a client's quarterly review can be checked against it line by line, and an auditor can trace it back to the records it came from.
Updating a figure without losing its history
- Set the period and the date of the next update, and state both on the page: "next update [month year]".
- Compute each figure from the same records, the same way, every time. If the method changes, say what changed and start a new series.
- Keep the earlier periods beside the new one, so a reader sees the trend and a weaker quarter is not quietly dropped.
- Change the page's date only when a figure changes. Google's guidance on people-first content lists changing a page's date to make it seem fresh, when the content has not substantially changed, among its warning signs.
- On the same day, update every copy: the service pages, the structured data and any directory listing that quotes the figure.
A dated figure is also what lets you correct an answer that quotes an old one. The article on reading answers by lane adds a column for the capacity an answer states, beside your current dated figure, so an out-of-date number shows up as a task with a page attached. That is freshness you can show, not a date you moved.
What not to publish
- A comparison with an industry benchmark or a competitor's figure. You would be quoting a number whose base and method you cannot show.
- "Industry-leading" or "best in class" beside a figure, which turns a measurement back into a claim.
- Your best month as the headline, with the other months missing.
- A rate with no count behind it, so nobody can tell whether it describes [n] shipments or [n] thousand.
- A figure from a site you have since left or a service you no longer run.
For a fulfillment center, the cut-offs and pick-and-pack figures a founder asks about belong on the public pages rather than in the client portal, as the article on public pages and the portal explains. The overview for logistics providers lists the other facts a logistics site has to state.
Common questions
Our on-time rate dropped last quarter. Should we still publish it?
Yes, beside the quarters before it and with the reason if you know it. A series with a dip and a recovery tells a buyer more than a single best number, and a buyer who audits you can ask for the records anyway. Dropping a bad quarter from the page, or changing the definition to lift it, is the kind of change a careful reader notices.
What counts as on time?
Whatever you define, provided the definition sits next to the number: the date you measure against, the event that counts, the tolerance, the unit and the exclusions. Pick the definition your contracts use, so the published figure and the one in a client's review agree. Change it only with a note saying what changed, and start a new series when you do.
Should we publish cut-offs for every site?
For every site that ships or receives, yes, on that site's page and with its time zone. A single cut-off on the homepage implies every building works to the same clock. If several sites do share one cut-off, say so once and link to it from each site's page, so a buyer never has to guess which building a time belongs to.
Can we compare our figures with an industry average?
Not on your own page. An average you did not measure comes with a source, a method and a base you cannot show, and quoting it borrows all three. Publish your own figures with their method. A buyer comparing providers will set them side by side, and the figure that carries its definition is the one that still means something there.
Sources
The one figure from elsewhere is the benchmark result in Aggarwal et al., cited for what that paper measured and nothing more; the page quotes no industry benchmark for logistics. The warning about page dates is from Google's own guidance, linked where it is used.
Sources: Aggarwal et al., GEO: Generative Engine Optimization, KDD 2024 · Google, creating helpful, reliable, people-first content
Read on