Container Schedule Reliability Explained: How It Is Measured and How Shippers Should Use It

Container schedule reliability is the percentage of vessel arrivals that land within a set tolerance of the schedule a carrier published. It is a statistic about ships, measured port call by port call, not a promise about your container. The most quoted series is Sea-Intelligence's Global Liner Performance (GLP) report, which also publishes a second figure, the average delay of the vessels that were late. Use these numbers to compare carriers and trade lanes and to size your buffer, but never read "62% reliable" as "62% of containers arrive on the booked date". Delays you see on the quayside also come from missed cut-offs and rolled cargo, which these indices largely do not capture.

What does schedule reliability mean in container shipping?

Every reliability figure has the same skeleton: take all the vessel arrivals in a period, compare each actual arrival with a scheduled arrival, count the ones inside a tolerance window, and divide. The differences are in three choices.

• The baseline schedule. The original published (proforma) schedule, or the latest schedule the carrier announced before arrival. The first punishes revisions; the second forgives them.
• The tolerance window. Typically about one day either side, but providers state it differently.
• The scope. All port calls on a service, or only the destination port of a trade. A call in the origin region is easier to make on time than a call at the end of a long rotation.

Because of these choices, two serious providers can publish figures for the same month that differ by double digits, and both are internally consistent.

How does Sea-Intelligence measure schedule reliability?

Sea-Intelligence's monthly GLP report, described on its own services page (read 27 Oct 2026) as covering more than 12,000 vessel arrivals, 60 or more container carriers and 34 trade lanes, is the series most often quoted in the trade press. Each monthly press release gives two headline numbers: the global schedule reliability percentage and the average delay for late vessel arrivals, in days. Two examples from its own releases:

• Release of 24 February 2026: January 2026 global reliability was 62.4%, and the average delay for late vessels was 5.17 days.
• Release of 26 January 2026: the December 2025 late-vessel delay was 5.04 days.

The two numbers answer different questions. Reliability says how often a vessel was on time. Average delay says how bad it was when it was not. A service can score a modest percentage but rarely be more than a day late, or score well and still deliver the occasional week-long slip. You need both.

A caution on the definition. The press releases I could read do not restate the on-time tolerance, and the full methodology sits in the paid report. Secondary descriptions of the GLP method say a vessel counts as on time if it arrives within one day of schedule, but they disagree on whether that is calendar or business days, so treat the exact rule as something to confirm from the report itself. The releases also report alliance results in two ways: "All arrivals" (including origin-region calls) and "Trade arrivals" (destination ports only), which matters when you compare alliances to older figures.

Why do different providers publish different reliability numbers?

Because the baseline and window differ. Two published methodologies show the spread, and both gave January 2026 readings, so the gap is easy to see.

Do not convert between series or average them. A schedule frozen two weeks out will look better than one measured against the original proforma, and a destination-only test is harder than an all-calls test. Pick one series, check its definition and read its trend, not its level.

As far as I could confirm, the Digital Container Shipping Association (DCSA) does not publish a reliability league table. Its role is data standards: its Operational Vessel Schedules standard defines planned, estimated and actual arrival and departure timestamps so that carriers and tools exchange the same fields. Those timestamps are what a reliability calculation is built from, but the on-time rule is each analyst's own. I could not confirm a published methodology for Linerlytica's carrier performance figures, so I do not describe it.

Reliability per port call versus per container

Reliability series are measured per vessel arrival. Your shipment lives on a different clock, and the vessel being on time is only one link in it:

1. Booking and cut-off. If your box misses the CY, SI or VGM cut-off, the vessel can sail on time without you. The rolled-cargo guide covers this, and it is invisible in a vessel index.
2. Vessel arrival. What the indices measure.
3. Discharge and availability. A vessel on time at the berth can still queue for a berth or yard space, and your container may come off after others.
4. Customs, gate-out and inland leg. Free time on demurrage and detention starts running from discharge, not from your warehouse being ready.

So "on time" means the ship arrived inside the window. It does not mean your container was discharged, released or delivered on the date you were told at booking. Transhipped cargo adds another vessel and another handover; see direct versus transhipped sailings.

Why did schedule reliability fall after 2020, and where is it now?

Trade press reports citing Sea-Intelligence put reliability at roughly 70% on average in the 2018 to 2020 period, falling to a series low of 32.0% in December 2021, the lowest since measurement began in 2011, with port congestion the main driver (reported by Maritime Executive and Seatrade Maritime, December 2021). Congestion at ports and inland, vessel queues and equipment imbalances pushed late-vessel delays above seven days in late 2021.

Reliability recovered but not to the old band. Sea-Intelligence's own releases show 62.8% for December 2025 (release of 26 Jan 2026) and 62.6% for June 2026 (release of 27 Jul 2026). Detours around the Cape of Good Hope, bunching of vessels at the end of a long loop, and congestion in Asian ports kept the level well under the pre-2020 norm. In its release of 29 September 2026 the series dropped sharply again for August; the reading and context are in our early October market update. Check Sea-Intelligence's press room for the current month before quoting any number, as these figures move every month.

How do alliances and blank sailings affect reliability?

Carriers sail in alliances that share vessels on a loop, so reliability is often a property of the service rather than the brand on your booking. The same press releases rank alliances separately from carriers; see who sails with whom in 2026 for the current structure. Blank sailings matter in a different way: a cancelled sailing is not a late arrival, so many series either exclude it or count it only if announced late. Kuehne+Nagel's method, for example, leaves out blank sailings and omitted calls announced after the 14-day freeze. An index can therefore look healthy while the next departure on your lane was cancelled and your cargo waited a week. Read blank sailings and void sailings explained and ask for the sailing history, not only the percentage.

How many buffer days should you plan?

There is no standard number, and a generic one would be wrong for your lane. Build it from data instead:

1. List the last 10 to 20 sailings of the specific service you use (not the carrier average) with the ETA you were quoted at booking and the actual discharge date.
2. Measure the slip of each, from the booking ETA to the actual date, including any rolls and omitted calls.
3. Choose a service level. Sort the slips and pick the one you would accept being exceeded rarely, such as the second-worst in twenty. A line item that stops a production run needs a higher cut than restockable retail goods.
4. Add the non-vessel steps: cut-off margin on the origin side, expected port dwell, customs release and inland leg.
5. Reset it each quarter. A congestion episode or a change in the service rotation makes last season's history obsolete.

The average late-vessel delay in the GLP releases is a sensible sanity check for the scale of a bad case (a little over five days in the 2025 and early 2026 readings cited above), but it is a global average, not your lane's.

What to ask a forwarder about the reliability of a specific service

• Which service (loop name and vessel rotation) is this quote on, and is it direct or transhipped?
• What was this service's on-time record over the last three months, and against which baseline: original schedule or latest announcement?
• How many sailings on it were blanked, omitted or rolled in the same period?
• What is the cut-off for this booking, and what is the plan if it is missed?
• Which alliance partner operates the vessel, and does the booking move between partners?
• What are the free time terms at the destination, so a late discharge does not become a detention charge?

What to do next

1. Choose one reliability series, read its stated definition and track its trend monthly rather than comparing levels across providers.
2. Ask your forwarder for the sailing history of the exact service you book, not a carrier-wide average.
3. Build a buffer from your own lane's slip data and revisit it every quarter. For route-level planning, see ocean freight transit times by route.
4. Protect the controllable steps: hit the cut-offs and document readiness so the only delay left is the one you cannot influence.

More than 29,300 logistics companies are searchable by country and service in the CargoLinked directory, and the public requests board lists freight that shippers have posted for forwarders to quote on directly. When you compare quotes, compare the service behind each one as well as the price.

This article is general information, not legal, customs or financial advice. Reliability figures change monthly and providers define them differently; confirm current numbers and definitions with the publisher and your forwarder or carrier before relying on them.