Delay Repay claims rejected

[Taunton]

I had one from 04.12.04 rejected:



It really should not be so difficult.

This is a system called Artificial Intelligence. Which I call Artificial Stupidity.

 

[AdamWW]

TfW have a nice feature in their app where you can make a delay repay claim for a journey in your pay as you go history...

...except it's just a link to their general delay repay process where you have to construct the claim from scratch.

At least it does list the PAYG journeys to use in lieu of uploading a ticket.

 

[SN_OCC]

This is a system called Artificial Intelligence. Which I call Artificial Stupidity.

Or ARS for short..........

 

[JordR]

This is a system called Artificial Intelligence. Which I call Artificial Stupidity.

I think it's acceptable to use automated systems that compare timetable and running data in the first instance. This has both resource benefits for the TOC and speed benefits for the majority of passengers, I would imagine you could clear >80% of claims this way?

Maybe what would be a better customer experience is for automated rejections to be more in the form of "our automated checks determined X so we are unable to make an automated payment, is there any information you would like to provide so a member of our team can consider your case further?".

 

[ainsworth74]

Maybe what would be a better customer experience is for automated rejections to be more in the form of "our automated checks determined X so we are unable to make an automated payment, is there any information you would like to provide so a member of our team can consider your case further?".

And then make it possible to actually give them a written explanation of what's gone wrong along with having it reviewed by someone who is actually willing to read and look into what has been said....

 

[robbeech]

I’ve long suggested that automated systems are fantastic at dealing with most claims but should work on an Approve or refer for manual checks process rather than an approve or reject for a reason pulled out of a hat process.

Of course, the current setup means thousands and thousands of people will give up. So spending money to cost money isn’t in their interests.

As such it really ought to be regulated but as the railway pays no attention to current regulation it’s unlikely it would do so for any more.

 

[SuspectUsual]

I think it's acceptable to use automated systems that compare timetable and running data in the first instance. This has both resource benefits for the TOC and speed benefits for the majority of passengers, I would imagine you could clear >80% of claims this way?

Maybe what would be a better customer experience is for automated rejections to be more in the form of "our automated checks determined X so we are unable to make an automated payment, is there any information you would like to provide so a member of our team can consider your case further?".

What would be far more customer focused would be having the automated system approve those that it can (your 80%), rejected those it knows are wrong (eg delay didn’t reach threshold) and refer the remainder to a human being with advice to the customer

 

[Fawkes Cat]

What would be far more customer focused would be having the automated system approve those that it can (your 80%), rejected those it knows are wrong (eg delay didn’t reach threshold) and refer the remainder to a human being with advice to the customer

Cases 2 and 3 here aren't mutually exclusive: if the threshold is timed to signals outside the platform, or if there was a delay in opening the doors, then a passenger may think they have a legitimate claim even if the system's automated metrics disagree.

 

[SuspectUsual]

Cases 2 and 3 here aren't mutually exclusive: if the threshold is timed to signals outside the platform, or if there was a delay in opening the doors, then a passenger may think they have a legitimate claim even if the system's automated metrics disagree.

That’s true. But the broader point is that the current system is configured as “if in doubt reject, let the customer do the legwork” and I think that’s wrong

 

[JordR]

You could potentially put a margin in automated rejections making it more likely to be correct on time even if the delay is measured based on a timing point that is not quite at the platform, say <10 minutes or <25 minutes delayed depending on the scheme.

I think there's a fair case for passengers and TOCs to automate what they can but the bit that needs to be fixed is that there should be a small proportion of cases that are currently being automatically rejected that should probably be manually reviewed.

It's also fair to say that without effective regulatory enforcement most commercial organisations will not invest in something that costs them money. In many industries a customer satisfaction improvement or repeat purchase improvement would contribute to such business cases, but from what I read elsewhere here that seems unlikely in the railways. The publicly owned TOCs still seem to act quasi-commercially, maybe GBR will eventually take a different centralised approach on delay repay?

 

[wilsontown]

Here's one from TPE:

The intended leg from 17:42 LEEDS, scheduled to arrive at YORK at 18:07 could not be taken because the intended journey was retimetabled. The alternative leg from 17:46 LEEDS left on time, scheduled to arrive at YORK at 18:19, actually arrived 18:21. Total journey delay: 14 minutes - your intended arrival time was 18:07 and the calculated arrival time for your journey was 18:21.

So they're saying that instead of waiting for the intended 1742 (which was delayed rather than "retimetabled"), I should have got the 1746 Northern stopper instead. But as the 1746 calls at Ulleskelf I expected that the 1742 would be able to overtake it and would arrive at York first. In the event the 1742 was routed behind the 1746 through Ulleskelf and eventually got to York 16 minutes late. It seems a bit harsh to expect us to guess which train will arrive first...

Not sure if it's worth bothering for the sake of a couple of quid but I might try to get them to reconsider.

 

[bakerstreet]

Here's one from TPE:



So they're saying that instead of waiting for the intended 1742 (which was delayed rather than "retimetabled"), I should have got the 1746 Northern stopper instead. But as the 1746 calls at Ulleskelf I expected that the 1742 would be able to overtake it and would arrive at York first. In the event the 1742 was routed behind the 1746 through Ulleskelf and eventually got to York 16 minutes late. It seems a bit harsh to expect us to guess which train will arrive first...

Not sure if it's worth bothering for the sake of a couple of quid but I might try to get them to reconsider.

I was under the impression that delay repay was intended to reflect what actually happened in the real world, rather than what could have happened had different decisions been made by the passenger. Especially when this isn’t a simple case of one operator and one line and one stopping pattern.
This decision seems particularly unfair, of course, because the alternative has been provided to you with the benefit of hindsight.

 

[SuspectUsual]

I was under the impression that delay repay was intended to reflect what actually happened in the real world, rather than what could have happened had different decisions been made by the passenger

It is, as long as the passenger doesn’t go rogue and deliberately delay themselves

 

[AdamWW]

It is, as long as the passenger doesn’t go rogue and deliberately delay themselves

Not very nice.

 

[Brent Goose]

It is, as long as the passenger doesn’t go rogue and deliberately delay themselves

I was under the impression that delay repay was intended to reflect what actually happened in the real world, rather than what could have happened had different decisions been made by the passenger. Especially when this isn’t a simple case of one operator and one line and one stopping pattern.
This decision seems particularly unfair, of course, because the alternative has been provided to you with the benefit of hindsight.

I have had the 20/20 hindsight before, sometimes it takes a couple of goes before reality sets in

 

[JordR]

I've certainly seen automated delay repay systems prompt me with several alternative delayed journeys to select from, so perhaps some acknowledgement that a passenger may not be able to get the second best option perfectly correct when adjusting to a delay.

 

[AdamWW]

I've certainly seen automated delay repay systems prompt me with several alternative delayed journeys to select from, so perhaps some acknowledgement that a passenger may not be able to get the second best option perfectly correct when adjusting to a delay.

I'm not sure I've ever been asked which trains I did catch.

Just which one I meant to catch and what I think the delay was (and I don't know why they ask that because they work it out for themselves anyway).

 

[Salteena]

Glad I found this very long-running discussion after racking my brains for some time over a disputed Delay Repay from WMR. Apparently 'We have been unable to approve your claim as the ticket you have supplied is not valid for the journey you have entered'.
I can’t understand this wording at all. It was a split ticket, one change, the delay meant I missed the connection (along with about 20 other people, so they may be turning down quite a few claims).
I submitted screenshots of both e-tickets and had been thinking it was much easier than faffing about with taking photos of defaced paper tickets.
I've now had to escalate it because I just can’t see what’s wrong…ugh.

 

[wilsontown]

Here's one from TPE:



So they're saying that instead of waiting for the intended 1742 (which was delayed rather than "retimetabled"), I should have got the 1746 Northern stopper instead. But as the 1746 calls at Ulleskelf I expected that the 1742 would be able to overtake it and would arrive at York first. In the event the 1742 was routed behind the 1746 through Ulleskelf and eventually got to York 16 minutes late. It seems a bit harsh to expect us to guess which train will arrive first...

Not sure if it's worth bothering for the sake of a couple of quid but I might try to get them to reconsider.

Well, I appealed this and they decided to pay up after all. Everyone involved has spent way too much time over the £2.41 that was at issue, but there it is...

 

[RyanOPlasty]

Customer satisfaction would be greatly improved if payouts were made for actual delays of more than 28 or 58 minutes whilst maintaining the official 30 and 60 minutes in documentation. Pepole would not then feel conned by petty arguments about the actual delay.

 

[AdamWW]

Oh dear.

Just had an email "approving" a Delay Repay claim for TfW, except they've paid out for 15 minutes not 30 because (having caught the train after the one I wanted to because that one was cancelled) I arrived 1 minute early.
Though in fact on time by the working timetable. And if the train they cancelled had run on time it would also have been "early".

And in any case they clearly say they pay for delays based on when you were "supposed" to arrive, not when the timetable is fiddled to say so.

Well none of this was a surprise, but this bit was:

Just to make it more fun, the link to make an appeal in the email approving the claim fails because the claim hasn't been approved yet.

The way it works with the new(ish) PAYG is a bit curious.

Make a claim for a journey early in the week before you hit the cap and the payment is based on the single PAYG price.
Do it once you've hit the cap and it's based on 1/20th the cap.

I suppose the consistent way would be to make people wait until the end of the week to make a claim.

 

[Kite159]

Customer satisfaction would be greatly improved if payouts were made for actual delays of more than 28 or 58 minutes whilst maintaining the official 30 and 60 minutes in documentation. Pepole would not then feel conned by petty arguments about the actual delay.

Agreed, make the thresholds 14/29/59 minutes whilst maintaining the offical 15/30/60 minutes in public facing documents, which will take into account those occasions where a train arrives 30 seconds early. Customers won't feel so annoyed when their hourly service only gets them a 30-59 minute payout because "you were only delayed by 59 minutes and 45 seconds".

 

[Belperpete]

Agreed, make the thresholds 14/29/59 minutes whilst maintaining the offical 15/30/60 minutes in public facing documents, which will take into account those occasions where a train arrives 30 seconds early. Customers won't feel so annoyed when their hourly service only gets them a 30-59 minute payout because "you were only delayed by 59 minutes and 45 seconds".

The reality in today's world is that it would soon become public knowledge that the thresholds were actually 14/29/59.

 

[robbeech]

The reality in today's world is that it would soon become public knowledge that the thresholds were actually 14/29/59.

But this doesn’t make a difference, if they advertised it as 14 29 59 you’d get people wanting it to be 13 28 58 hence the need to not change any documented times.


Are there any berthing offsets that are too far the other way? For example is there anywhere that has a 1 minute offset but it consistently takes 15 to 20 seconds?


Or are they all wrong in the same (very convenient for the railway’s delay repay policy) way?

We never hear about it the other way but is that because there’s no reason to hear about it as it doesn’t affect passengers or is it because trains are generally slower now into stations due to increased traffic, or something else?

 

[35B]

Agreed, make the thresholds 14/29/59 minutes whilst maintaining the offical 15/30/60 minutes in public facing documents, which will take into account those occasions where a train arrives 30 seconds early. Customers won't feel so annoyed when their hourly service only gets them a 30-59 minute payout because "you were only delayed by 59 minutes and 45 seconds".

That’s about sensible use of discretion on edge cases, and the human hatred of falling just the wrong side of a threshold.

 

[AdamWW]

That’s about sensible use of discretion on edge cases, and the human hatred of falling just the wrong side of a threshold.

But it's made much worse by the fact that many timetables are clockface so the threshold has been put exactly where lots of journeys will run foul of it when a train is cancelled, and where people wouldn't even expect a threshold problem to apply. Your train was cancelled and you got the one an hour later- what's the delay? The common sense (and in my view correct given how the delay repay rules are phrased) answer is one hour.

And if it's so reasonable to use padding in timetables and resulting "early" arrivals to justify paying for a 30 minute delay when an hourly train is cancelled, why not say so explicitly in the FAQs? Why not explain that the definition of delay is based on comparing apple and oranges - actual arrival of replacement train vs (in many cases padded) passenger timetable arrival time of the one that was cancelled?

Instead they just mutter about standard industry processes (i.e. we're all at it).

 

[35B]

But it's made much worse by the fact that many timetables are clockface so the threshold has been put exactly where lots of journeys will run foul of it when a train is cancelled, and where people wouldn't even expect a threshold problem to apply. Your train was cancelled and you got the one an hour later- what's the delay? The common sense (and in my view correct given how the delay repay rules are phrased) answer is one hour.

And if it's so reasonable to use padding in timetables and resulting "early" arrivals to justify paying for a 30 minute delay when an hourly train is cancelled, why not say so explicitly in the FAQs? Why not explain that the definition of delay is based on comparing apple and oranges - actual arrival of replacement train vs (in many cases padded) passenger timetable arrival time of the one that was cancelled?

Instead they just mutter about standard industry processes (i.e. we're all at it).

I agree. I’ve many times been running half an hour late and prayed for a delay outside the terminal station so that the padding would be counteracted. Or another situation where I could be guaranteed that if my connecting train was 15 late, I’d get in 61 late and have a valid claim.

That doesn’t alter the basic point. The public timetable is part of what the railway offer me, and Delay Repay is measured against that. Any threshold will have challenges because people feel hard done by by being just short, but the answer there is about responding appropriately to the real world impact, not fiddling with the rules.

 

[SN_OCC]

I had a claim rejected some months ago because the train arrived 14 3/4 late minutes...............

Technically it was outside the threshold. But I still miss my connection

 

[Haywain]

But I still miss my connection

In which case it is the arrival at your final destination which determines your eligibility for Delay Repay compensation.

 

[SN_OCC]

In which case it is the arrival at your final destination which determines your eligibility for Delay Repay compensation.

Agreed. Unfortunately I am now stuck in a perpetual loop of arguments between two train operators who each blame each other and say it is the other operators responsibility. Its is a bit like arguing with my dogs........totally pointless