If the carrier mix on your Q4 lanes was set against performance figures from earlier this year, the buffer you are carrying may now be wrong in both directions. August put 45 percentage points between the best and worst performers among the twelve largest carriers, while average carrier delays ranged from 2.8 days to more than 10 days.
Xeneta’s August 2026 Schedule Reliability Scorecard, published on 17 September 2026, is the dataset behind that spread. This post covers what the numbers say, where the gap between carriers actually comes from, and what it changes for buffer setting ahead of Q4 tenders.
August in one table
| Metric | August 2026 | Change |
|---|---|---|
| Global on-time arrivals | 29% | Down 4pp from July |
| Average delay | 5.1 days | Up from 4.2 days in July |
| Consecutive monthly declines | 3 | June 37%, July 33%, August 29% |
That 29% is the second-worst reading of the past twelve months and erases the recovery built off February’s floor of 27%. Xeneta’s own historical marker is worth holding onto: the last time global on-time arrivals sat at or below 29% for a sustained period was July 2024 through January 2025, during the closing months of THE Alliance and 2M.
Frequency and severity moved together, which is the part that makes recovery harder. More ships arrived late and the late ones arrived later.
The trades did not move together
| Trade | On-time, August 2026 | Change |
|---|---|---|
| Far East to Europe | 6% | Down 10pp |
| Africa | 21% | Down 15pp |
| South America East Coast | 34% | Down 12pp |
| South America West Coast | 37% | Top 3 most reliable |
Far East to Europe is the headline, at 6% on-time with 8.2 days of average delay. Xeneta places that at levels last seen at the height of the pandemic in late 2021. The contributing factors have been accumulating since the Red Sea crisis began, through Cape of Good Hope routings, longer berth stays for larger vessels, North Europe port congestion and repeated extreme weather.
South America East Coast is the useful counter-example, because the cause is identifiable and short. A three-day pilotage strike across Argentina’s ports at the start of August pushed average berth arrival delays at ports including Buenos Aires up by 1.6 days and encouraged carriers to divert volumes to Montevideo and Itajaí. Average delay on the trade doubled to 5.4 days. The timing landed on the last month of Argentina’s primary export season, running into Brazil’s peak.
Forty-five points between carriers in the same market
Ten of the top twelve carriers declined in August and six hit twelve-month lows: Hapag-Lloyd, CMA CGM, PIL, HMM, Wan Hai and Yang Ming. The spread between best and worst narrowed from 48 percentage points in July to 45 in August, and it narrowed for the wrong reason. The leaders fell rather than the laggards catching up.
| Carrier | On-time, August 2026 | Change | Average delay |
|---|---|---|---|
| Maersk | 51% | Down 6pp | 2.8 days |
| Hapag-Lloyd | 46% | Down 7pp | 2.8 days |
| CMA CGM | 33% | Down 7pp | 3.7 days |
| Cosco | 22% | Down 6pp | 5.7 days |
At the bottom of the ranking the numbers separate further. HMM recorded 9% on-time performance with 9.1 days of average delay, while Yang Ming recorded 6% with 10.1 days. Against Maersk’s 51%, Yang Ming’s result creates the full 45-percentage-point gap between the strongest and weakest carrier in the ranking.
Set those two ends side by side and the planning consequence is direct. Market-wide carrier averages ranging from 2.8 days of delay to more than 10 days point to materially different operating risk. They do not tell you what will happen on a specific port pair, which is why carrier and service performance needs to be checked on the lanes you actually use.
One more thing changed in August that undercuts a common shortcut. Hapag-Lloyd’s 46% is its lowest on-time performance since before Gemini Cooperation launched in February 2025, and Maersk gave up 6 points in the same month. Buying the alliance instead of buying the service no longer does the work it did earlier this year, and the Gemini network itself is still moving, with Maersk and Hapag-Lloyd announcing in September that four more services will transition back through Suez. Those routing changes reset transit times against which earlier reliability history was built.
Why two providers give you two different numbers
Anyone checking more than one reliability source this year will have noticed the readings do not line up, and the gap is not small. For June 2026, Sea-Intelligence put global liner schedule reliability at 62.6%, while Xeneta recorded 37%.
The difference is methodology. Xeneta measures arrivals against the originally published carrier schedule. Sea-Intelligence also measures actual vessel arrivals against scheduled arrivals, but uses a different schedule baseline, coverage and aggregation methodology. The resulting figures are internally useful within each dataset but should not be compared as though they measure exactly the same thing.
For an operations team, the practical rule is simpler than the methodological detail. Use one provider consistently when comparing carriers or tracking changes over time, and do not transfer an absolute reliability percentage from one dataset into another.
Two rules follow. Keep one provider inside any single buffer calculation, and never compare one carrier’s figure from one provider against another carrier’s figure from a second.
Where the delay is being manufactured
| Region | Average vessel wait, August 2026 |
|---|---|
| Africa | 3.3 days |
| Middle East | 2.3 days |
| Southeast Asia | 1.7 days |
| Northeast Asia | 1.3 days |
Africa held the top position for a second month on structural grounds rather than event-driven ones. Beira averaged 11.8 days of wait, and that average rests on 17 vessels waiting per day rather than one extreme outlier.
The Middle East spreads its congestion thinly instead of concentrating it. Sharjah ran 4.0 days, Djibouti 3.9 and Sohar 3.8, while the region’s largest gateway at Jeddah sat at 0.6 days. A routing decision made on a regional average would be wrong at almost every individual port in it.
Northeast Asia carries the trap worth internalising. It ranked fifth on wait time at 1.3 days while holding the third-highest congestion ratio, with 16% of vessels in a holding pattern on any given day. Ningbo makes the mechanism visible: average wait actually improved slightly to 2.8 days from 3.0 in July, while vessels at anchorage in Ningbo-Zhoushan rose from 39 per day to 63, an increase of 63% month on month. The wait-time metric fell while the queue behind it grew by nearly two thirds. This is the same pattern behind Shanghai’s slide to 21% schedule reliability earlier in September.
Those queues do not stay in Asia. Four back-to-back typhoons left 1.1 million TEU laid up at anchorage across Ningbo, Shanghai and Yantian in the week after the final storm passed, and vessels accumulating at Chinese anchorages in August convert into berth arrival delays in Europe and North America over the following weeks. Mid-Autumn Festival on 25 to 27 September and Golden Week from 1 to 7 October arrive back to back this year, carrying ten days of factory closures and compressing bookings into the weeks before them.
If you are carrying one buffer across every carrier on a lane because per-carrier arrival performance is not something your team can see in one place, walk through how ops teams track arrival performance and exceptions across carriers.
What this changes for Q4 planning
Buffer per carrier per port pair, not per lane. Market-wide performance shows that carrier reliability can diverge sharply, but the buffer itself should come from performance on the specific carrier, service and port pair you use. A single lane buffer can hide that variation and produce a number that is too thin for weaker services and too generous for stronger ones.
Read on-time rate and delay severity as two separate variables. A carrier at 46% on-time with 2.8 days of average delay creates a different operational problem from one at 9% with 9.1 days. The first misses the window and lands within the week. The second can break the week.
Re-check any carrier mix that was set on first-half figures. Six of the top twelve reached twelve-month lows in a single month, which is enough movement to invalidate a ranking built in Q2.
Track queue length alongside wait time at your discharge ports. Ningbo in August is the worked example of a wait-time average falling while exposure grew.
What to re-check before your next tender round
- Pull your own on-time performance by carrier and by port pair for the last three months, rather than by lane.
- Separate the on-time percentage from average delay for each carrier, and set the buffer off the delay figure.
- Confirm which provider’s methodology sits behind any reliability figure quoted in a tender document, and keep one provider throughout.
- Identify which of your services have changed routing since the reliability history you are quoting was recorded.
- For Far East to Europe volume, plan against arrival variance rather than an on-time rate, since 6% on-time gives an on-time figure almost no planning value.
- Check queue length as well as vessel wait time at your top three discharge ports.
Further Reading
- Xeneta: Schedule Reliability Scorecard, August 2026
- Seatrade Maritime: Shippers unimpressed by container line reliability
- Xeneta: Schedule Reliability Scorecard, Q2 2026
Sourcing: all global, trade, carrier and port congestion figures are from Xeneta’s Schedule Reliability Scorecard for August 2026, published 17 September 2026, and describe August 2026 performance. The Sea-Intelligence comparison uses its June 2026 global reliability figure alongside Xeneta’s June 2026 figure. Reliability measurement methodology differs between providers, and figures from different providers are not directly comparable. Confirm carrier performance on your own port pairs before using any market-level figure in a tender or buffer calculation.
Need help interpreting this disruption or your shipment?
For a quick question, chat with Tradlinx on WhatsApp. For a deeper discussion, book a time below.
Prefer email? Contact us directly at min.so@tradlinx.com (Americas), sondre.lyndon@tradlinx.com (Europe), or henry.jo@tradlinx.com (EMEA/Asia).
Leave a Reply