◢The Payment On Hold payment on hold Open the partner account
paid
Affiliate disclosure. The partner link in the masthead and in the band beside the copy on this page is a sponsored link to a partner operator, and this site may be paid if you open an account through it, at no extra cost to you. It carries rel="sponsored noopener" and opens in a new tab. A page about a payment an operator held or refused should not leave its own funding unsaid: one link funds the site, no operator is named, rated or recommended anywhere on it, and no operator is described by any figure on it.
The Payment On Hold / clock
how long a stop lasts

How long a payment stays stopped

A stop has a mean, a median and a tail, and the tail is where most of the frustration lives. This page gives the three, separates the clock from the delay, and shows the 14 requests that sat past ninety days.

Desk spec
weighted mean
8.7 d
released median
6 d
nine in ten
34 d
past 90 days
14
same day
1,421
reversed mean
3.1 d
the stateWhat the request is doing. Of 1,800 sampled requests, 1,421 were paid without ever stopping, 380 were held at least once, 260 of those were released and 120 were closed without payment.
the groundThe recorded reason a payment stopped. Across the 380 stops: a document gap 148, a source-of-funds query 96, a rail or name mismatch 61, an alleged term breach 49 and a dependency not settled 26.
the retentionThe one outcome that costs money permanently. Of 742,500.00 requested, 693,412.50 was paid, 41,147.30 returned to balances and 7,940.20 retained, at a mean 214.60 per reclaimed request.
Direct answer

A stop lasts a weighted mean of 8.7 days across 380 sampled stops, but the median released request is back in six days, and nine requests in ten are paid or closed within 34 days. Fourteen sat past ninety days, which is 0.8% of all requests and 3.7% of the stopped ones.

Three numbers, three uses

The mean answers what the sample cost in total; the median answers what a typical request experiences; the ninetieth percentile answers what the worst ordinary case looks like. Reporting only the mean would make a stop look worse than it usually is, because the alleged-breach ground pulls it up, and reporting only the median would hide the tail entirely.

the clock, by ground
groundstopsmean daysdays x stops
a document gap1484.2621.6
a source-of-funds query9611.61,113.6
a rail or name mismatch616.8414.8
an alleged term breach4922.41,097.6
a dependency not settled262.154.6
all stops3808.73,302.2
the tail, stated honestly 1,800 requests paid the same day 1,421 (78.9%) resolved inside 34 days 1,786 (99.2% of requests) past 90 days 14 (0.8% of requests, 3.7% of stops) of the 14: 9 were alleged-breach stops, 4 source-of-funds, 1 a rail mismatch mean of the 14 141 days

The clock against the delay

The clock is the time the stop lasted; the delay is how much later the money arrived than it would have without the stop. They are equal only when the request would otherwise have been paid the same day. Because 1,421 of the sample were paid the same day, the base case is same-day, so for the 260 released requests the delay is the whole clock. For the 22 that were refused and then reversed on appeal, the delay runs from the request to the reversal, which is a mean 7.9 days rather than the 3.1 days the appeal itself took.

What the re-ask does to the clock

A stop that needs two items is not simply twice a stop that needs one: each item restarts the operator reading time. In the sample a released request needing one item closed in a mean 4.8 days and one needing four or more in a mean 12.4 days, so the last three items cost about three times as much time as the first one.

Judging whether a stop has run too long
  • Compare the stop against its ground, not against a general rule: 11.6 days is normal for a source-of-funds query and unusual for a document gap.
  • Count the items you have supplied; each additional item restarts a reading time.
  • Treat 34 days as the ordinary outer edge, because nine requests in ten closed inside it.
  • Treat past 90 days as a tail event, present in 14 of 1,800 requests.
  • Separate the clock from the delay by asking when the money would have arrived without the stop.