Trainline/GWR anti-fraud partnership

[100andthirty]

This press release was posted this morning to rail trade publications:

'LONDON, 28th July 2021: Trainline has partnered with Great Western Railway (GWR) to reduce refund fraud by 39% across all its routes.

Trainline’s fraud and tech teams work together to identify customers most likely to claim a fraudulent self-serve refund through analysis of a combination of factors, including previous behaviour. Those most likely to have submitted a fraudulent claim are then systematically contacted in order to prevent further refund abuse.

Collaborating with Ravelin, the leading fraud detection and payments authentication provider, Trainline created the bespoke Refund Protect solution, available to all rail operators, providing them with the insights and support to prevent the refund system from being abused.

Trainline has successfully reduced its online fraud by half since October 2020, including a significant reduction in card payment fraud across its own app and website. Card fraud is where people attempt to buy tickets using stolen details, typically purchased on the dark web. Trainline’s data capabilities have also been key in identifying customers who have bought tickets with a railcard discount applied, without owning a valid railcard.

Champa Magesh, President, Trainline Partner Solutions, comments: “We’re pleased that our technology is supporting our rail operating partners in preventing fraudulent refunds from being claimed. Working closely with our partners to minimise fraud means our teams can be focussed on driving innovation for the benefit of customers, and attracting more travellers to rail.”

Kevin Forrest, Head of Revenue Protection Development, GWR, comments: “GWR has been working closely with Trainline in order to prevent, deter, identify and pursue fraud and fraudsters who commit journey related fraud within the rail industry. Part of the work involves looking at the volume and impact of potentially dishonestly obtained refunds. This project is significantly helping GWR create and deliver an intelligence-led revenue protection function.”'

 

[MarlowDonkey]

Trainline’s data capabilities have also been key in identifying customers who have bought tickets with a railcard discount applied, without owning a valid railcard.

If they can do that, they should also be able to identify customers who have bought tickets with a discount applied, who own a valid railcard, but who travel without it.

 

[87 027]

And of course you don’t have to own the railcard at the time of ticket purchase - it’s the date of travel that matters!

Or is this a sneaky change to the rules?

 

[Fawkes Cat]

Trainline’s data capabilities have also been key in identifying customers who have bought tickets with a railcard discount applied, without owning a valid railcard.

If they can do that, they should also be able to identify customers who have bought tickets with a discount applied, who own a valid railcard, but who travel without it.

(my bold and italics)
How would Trainline do the part I've highlighted? I suppose that GWR might be able to do that on-train or at-station, but would they be realistically able to access Trainline data there and then? And what about customers who haven't bought from Trainline?

 

[Bletchleyite]

And of course you don’t have to own the railcard at the time of ticket purchase - it’s the date of travel that matters!

Or is this a sneaky change to the rules?

This is one that has shifted with the move to electronic sales. Previously most booking offices would not sell a discounted ticket without a Railcard being presented, which meant that you effectively needed validity at both points. But then again, back then discounts were not available on advance purchase type tickets, and so people wouldn't be buying discounted tickets well in advance to start with.

 

[MarlowDonkey]

(my bold and italics)
How would Trainline do the part I've highlighted?

In some manner flag the ticket that not only had a discount been applied, but also that eligibility for the discount had been validated. If if became a feature only available with Trainline, it might even attract business despite Trainline's other disadvantages. For occasional travellers it gives them protection against incorrectly buying a discounted ticket for travel after their card had expired.

 

[AlterEgo]

(my bold and italics)
How would Trainline do the part I've highlighted? I suppose that GWR might be able to do that on-train or at-station, but would they be realistically able to access Trainline data there and then? And what about customers who haven't bought from Trainline?

It may be the case that a prompt comes up to the member of staff to check the railcard.

 

[skyhigh]

It may be the case that a prompt comes up to the member of staff to check the railcard.

Already happens with our handsets. If any discount is applied, it prompts us to check the validity of it.

I suspect it works more along the lines of during a ticket inspection it's noticed no railcard is held, a report is made to GWR, GWR then ask Trainline who produce a list of tickets purchased with that discount (rather than trying to identify people using the railcard discount who done own one at point of purchase)

 

[SteveM70]

Trainline’s fraud and tech teams work together to identify customers most likely to claim a fraudulent self-serve refund through analysis of a combination of factors, including previous behaviour. Those most likely to have submitted a fraudulent claim are then systematically contacted in order to prevent further refund abuse.

Bit of a leap of faith there!

All sounds a bit GAish to me

 

[35B]

This is one that has shifted with the move to electronic sales. Previously most booking offices would not sell a discounted ticket without a Railcard being presented, which meant that you effectively needed validity at both points. But then again, back then discounts were not available on advance purchase type tickets, and so people wouldn't be buying discounted tickets well in advance to start with.

I recently purchased a ticket using a Family Railcard (bought in a hurry the day before son’s 16 birthday!). The previous one was recorded on my account with LNER and I was unable to book (a month before travel) without recording a card on my account.

 

[robbeech]

Bit of a leap of faith there!

All sounds a bit GAish to me

Perhaps, perhaps not.

It’s pretty clear that this sort of thing is a significant problem industry wide. Ie heard people tell eachother about this “cool trick” they’ve found, it’s been happening for years and I certainly encourage the retailers and operators to work together to minimise this and take the necessary action against people abusing it.

I think the difference here is that in order to trigger their suspicions here you have to buy a ticket, and then refund it. It doesn’t rely on punctuality of the service it just relies on you not encountering a ticket check (in human or barrier form) which is very common.

Personally I see someone with 50 tickets and 30 refunds as orders of magnitude more suspicious than someone with 50 journeys and 30 delay repay claims.

What IS strange is why the retailers actually give a damn about it. From what I can gather this sort of thing tends to happen on lower priced tickets for shorter journeys, generally longer distance journeys end up being between larger stations with barriers and/or will almost certainly have a ticket check on board one or more of the trains.

if someone buys a £50 ticket the retailer makes a couple of quid from the transaction (I’ve no idea of the actual figure). If they refund that (and there is no disruption) then the retailer keeps £10, but as far as I know there are no charges from the operators meaning that until a ticket price goes above a certain point (see above about its likelihood based on distance) the retailer should BENEFIT from a refund, unless I’m mistaken.

 

[CyrusWuff]

if someone buys a £50 ticket the retailer makes a couple of quid from the transaction (I’ve no idea of the actual figure). If they refund that (and there is no disruption) then the retailer keeps £10, but as far as I know there are no charges from the operators meaning that until a ticket price goes above a certain point (see above about its likelihood based on distance) the retailer should BENEFIT from a refund, unless I’m mistaken.

When a ticket is refunded, the original transaction is reversed, so the retailer loses their commission for having sold it and the TOCs lose the revenue they received through the ORCATS allocation. (A slight oversimplification, I suspect, given the Admin Fee muddies things a bit, but you get the idea).

 

[[.n]]

Trainline’s data capabilities have also been key in identifying customers who have bought tickets with a railcard discount applied, without owning a valid railcard.

I [used to] quite often buy tickets with a railcard I didn't have - because I was buying them for a friend etc.

(Having said that never from Trainline, because of their fees,)

 

[robbeech]

When a ticket is refunded, the original transaction is reversed, so the retailer loses their commission for having sold it and the TOCs lose the revenue they received through the ORCATS allocation. (A slight oversimplification, I suspect, given the Admin Fee muddies things a bit, but you get the idea).

As I thought, so for many tickets a £10 admin fee which is kept by the retailer may well be a higher price than the commission they’d have had from the ticket, which is what makes it strange that the retailers are pushing this.
Unless of course the retailers have been refunding all tickets and not having a system in place to check whether they’ve been scanned and they’ve been losing lots of money when tocs have demanded the revenue.

 

[Nova1]

Good. I know far too many people who use trainline and just refund their tickets once they are through.

 

[CrispyUK]

I wonder what the volume of refund fraud is across the network and whether Trainline are coming under any scrutiny from the industry for facilitating easy refunds of tickets that haven’t been clipped during the journey? I imagine people are far more likely to try it on if they just need to tap a few buttons in the app compared to attempting to blag a refund at the ticket office of a station they’ve just travelled to!

Guessing that the ‘Refund Protect’ service is something the TOC has to pay for, if so it’s an additional revenue stream for Trainline.

 

[Ralph Ayres]

I love the claim of reducing refund fraud by 39% rather than say "by over a third". They must be very confident that they can identify them so precisely.

 

[Paul Kelly]

As I thought, so for many tickets a £10 admin fee which is kept by the retailer may well be a higher price than the commission they’d have had from the ticket, which is what makes it strange that the retailers are pushing this.
Unless of course the retailers have been refunding all tickets and not having a system in place to check whether they’ve been scanned and they’ve been losing lots of money when tocs have demanded the revenue.

With Trainline the fee is now only £10 for tickets with a face value of £15 and over: https://support.thetrainline.com/en...s/78000000551-tiered-refund-and-exchange-fees

I think it's much more likely that people are refunding tickets that haven't been scanned - Trainline would be doing nothing wrong by facilitating this and as you say making a nice profit from the refund fees. But I'd be surprised if the TOCs aren't pushing back about it!

 

[Sleepy]

Someone has obviously noticed a spike in refunds on mobile tickets during Covid due to lack of ticket inspections onboard. I know of a TOC that has picked up the difference in the statistics of their DOO routes compared to guard operated ones since revenue duties resumed.

 

[Joe Paxton]

m-tickets - the type that require activation in an app - are rubbish, for many reasons. Platform and media agnostic e-tickets are far superior.

However, from the railway's point of view, this type of fraud does arguably make m-tickets a bit more of an attractive choice compared to e-tickets.

 

[skyhigh]

Refund fraud is a significant issue- a surprisingly high percentage of e-tickets are refunded. In the days before onboard scanning at our TOC, it was common to be shown a screenshot of a ticket, and if you asked to see it in the app, it'd show as 'refund requested'.

It's particularly an issue on fully unbarriered journeys but now scanning is onboard, new tricks have appeared such as apps that mimic a cracked screen so that the aztec code won't scan. Other tricks have been noticed, such as buying unusual splits that don't save any money that start/end at a station with barriers- e.g. Leeds-Cross Gates, and Cross Gates-York. The first ticket is scanned on the barrier. If it's checked onboard before Cross Gates, the first ticket is shown. No further checks and the second ticket for the majority of the journey is refunded. If the second ticket is scanned, it's not refunded.

I'm aware that TOCs are working with Trainline to combat this - if a ticket is scanned and a refund is attempted, they may choose to issue a refund (despite knowing it's been scanned) to allow a pattern of evidence to build up. An unusual amount of refunds may also be investigated. Having spoken to someone who is involved in this, I understand this is partly where the data processing comes into play by trying to pick up on the various 'tricks' used to allow patterns of behaviour to be flagged for further investigation. The DfT currently has revenue loss from this right in it's focus.

I suspect we'll shortly start getting threads on here from people who have been caught out.

 

[Paul Kelly]

I suspect we'll shortly start getting threads on here from people who have been caught out.

I think there've been a few already although it's not always clear what's going on:

Transport Investigations prosecuting me

Last month i was travelling from Leicester to Nuneaton on a CrossCountry Train. I had bought the ticket on Trainline which i rarely use. As normal i scanned my ticket to get through the gate at the Leicester train station and boarded the train. During the journey i had a look on the trainline...

www.railforums.co.uk

Simple misunderstanding saw man threatened with court over train journey he paid for

This isn't about me, I'm not seeking help on penalty fares, fines and other post-travel ticketing issues, so this matter doesn't fit into the Disputes & Prosecutions thread.... Steve Mueller booked tickets online to travel from Abergele & Pensarn station to Chester via the Trainline app. He...

www.railforums.co.uk

 

[jon0844]

Short faring has been rife with onboard checks not done until recently, and as I've said elsewhere, the people who are eventually caught will be in deep water when their whole Trainline history is examined.

I am sure TOCs will be looking at how many people are buying a ticket from one station to the next, and printing it many miles away just 30 or 40 minutes before they got to the destination.

To this day I'd say most, if not all, of the people who received letters from GA regarding Delay Repay claims had misclaimed at least once, and possibly many more, so I am sure this will be pretty effective and well targeted. There are probably hundreds, if not thousands, of easy to prove cases.

I do hope the Rail Forums servers are suitably upgraded for the inevitable posts in the coming months.

I'm sure Trainline wants to help to a) keep in the good books of the DfT and b) because they make more money from the correct tickets being sold, as well as not being refunded.

 

[Merseysider]

I love the claim of reducing refund fraud by 39% rather than say "by over a third". They must be very confident that they can identify them so precisely.

I’d be very wary about taking at face value any statistic the Trainline come up with.

They do have a history of shouting random numbers without evidence or context.

Trainline ad banned for being misleading

It said that on average 39% could be saved by buying tickets in advance at the firm's website, thetrainline.com
...
the watchdog banned the TV ad

 

[AlterEgo]

Maybe they’re reducing fraud by 39% because they’re selling tickets 39% cheaper than the walk up price?

 

[Kite159]

Short faring has been rife with onboard checks not done until recently, and as I've said elsewhere, the people who are eventually caught will be in deep water when their whole Trainline history is examined.

I am sure TOCs will be looking at how many people are buying a ticket from one station to the next, and printing it many miles away just 30 or 40 minutes before they got to the destination.

To this day I'd say most, if not all, of the people who received letters from GA regarding Delay Repay claims had misclaimed at least once, and possibly many more, so I am sure this will be pretty effective and well targeted. There are probably hundreds, if not thousands, of easy to prove cases.

I do hope the Rail Forums servers are suitably upgraded for the inevitable posts in the coming months.

I'm sure Trainline wants to help to a) keep in the good books of the DfT and b) because they make more money from the correct tickets being sold, as well as not being refunded.

Or program ticket gates at terminal stations to reject tickets from the previous couple stations to challenge anybody trying to use a ticket to see where it was purchased. Or even have staff at the last station before barriers giving passengers tokens to say they actually boarded there. Ie Edge Hill/Aston etc.

Especially on those lines which operate DOO or with multiple units without gangways where the guard has to stay in the rear unit (Northern) or those with more limited ticket checks (Crosscity line)

 

[jon0844]

Or program ticket gates at terminal stations to reject tickets from the previous couple stations to challenge anybody trying to use a ticket to see where it was purchased. Or even have staff at the last station before barriers giving passengers tokens to say they actually boarded there. Ie Edge Hill/Aston etc.

Especially on those lines which operate DOO or with multiple units without gangways where the guard has to stay in the rear unit (Northern) or those with more limited ticket checks (Crosscity line)

I believe that it is currently not possible to reject any barcode tickets for a manual check, hence why these are going to be ripe for fraudulent activity.

However, staff on gates and trains will be able to scan them with their devices, which is why a lot of people will be eventually caught out - and then find themselves in very hot water as it won't be one journey they're being done for.

 

[Paul Kelly]

I believe that it is currently not possible to reject any barcode tickets for a manual check, hence why these are going to be ripe for fraudulent activity.

I had one rejected at Reading last week. I assumed it was because I was using a Network Railcard, but then the attendant just let me through without asking to see my railcard.

new tricks have appeared such as apps that mimic a cracked screen so that the aztec code won't scan

Is it possible to manually scan the ticket in such a situation, by typing in the code shown next to the aztec code?

 

[skyhigh]

Is it possible to manually scan the ticket in such a situation, by typing in the code shown next to the aztec code?

Sorry, completely missed this. It's not possible on any scanners I've ever used, but it's a good point. I'll raise it and see if it's possible.

 

[Wallsendmag]

I have the ability to manually check a UTN (Unique Ticket Number) by typing it in .