Great Western Coffee Shop

I went on "Sales Training" half a lifetime ago.   I'll admit that I scoffed at the whole idea, and I'll admit that I said that everything we were taught was obvious ... but it was only obvious once pointed out.  And one of those things pointed out was that we should put ourselves in the customer's eye and talk with him about how our product is of benefit to him, rather than about the features we put into it.

A feature of train services is the number of them that arrive at their final destination on time (or if you like an alternative measure, their average lateness).   A benefit of train services is getting people where they want to go, and a measure that's not just an industry statistic should be the number of passengers who arrive at their final destination on time (or if you like an alternative measure, their average lateness).

Can you spot the difference?   Let me describe my journey today.

The 14:08 off Braford-on-Avon was six minutes late. It left Bristol Temple Meads about 15 minutes late, and pulled in to Newport also 15 minutes late - so that everyone "connecting" via Hereford to Chester watched their train leave as it pulled in to the platform.  Realtimetrains tells me that it arrived into Cardiff around 12 minutes late.

The 16:10 from Newport to Manchester ran to time ... took me to Crewe (not really where I wanted to be!)

The 18:56 from Crewe to Chester also ran to time ... but although the train was on time, I was 50 minutes late there. (*)

The 19:30 Chester to Liverpool also ran to time ... just a shame my schedule was for the 18:30. (*)

Rail industry measurement - 4 trains, total lateness 12 minutes. Average lateness 3 minutes.

Customer's measurement - 1 journey, arrived 60 minutes after scheduled. Average lateness 60 minutes

Discuss  ;D

(*) - I will admit to writing this story on the train in anticipation of what would happen, as a good illustration of how  inappropriate the common measure is for customer performance.   As it was, I had a chat with the train manager from Newport and approached the train manager of a direct Crewe - Liverpool service which I was given permission to use.   So I should really adjust my figures above and admit to being only about 25 minutes late into Lime Street.

This doesn't just apply to journeys involving connections or multiple operators. As far as I can see, whatever the train companies measure internally, the only information on performance which is routinely communicated to the public is the passengers' charter punctuality and reliability averaged over a group of services over a month (although FGW also gives a daily percentage on its website). I know I can look at individual trains in realtimetrains.co.uk if I really want to, but I usually have better things to do. In the case of the services I use (peak time Westbury to London and back), I can therefore find out the percentage of FGW trains designated as "High Speed Services" which arrived at their destination within 10 minutes of the booked arrival time. This average performance over a large number of services is, however, often very different from my personal experience on the trains I use each day.

In an office I used to work in, one of my slightly obsessive commuter colleagues who travelled on South Eastern trains used to calculate his "personal" punctuality figures for his daily commute. The result tended to be between 50% and 60% of the trains he used arriving within 5 minutes of booked time, rather than the published figure of ~90%. This could partly be attributed to the fact that he didn't travel to the final destination of the train - he might be 7 minutes late his personal destination, but the train would miraculously become "on time" at its final destination, apparently due to extra time being added into the timetable between the last two stops. I can't corroborate this, but it sounds plausible.

Regarding the customer experience relating to journeys involving connections, I guess that with the train running data collected by the railway, it would be theoretically possible for someone to construct a computer model which would calculate the average punctuality of every end-to-end permutation of connecting services by comparing an achieved timetable based on actual timings against a planned timetable. I don't know if the industry has done any research to find out what type of performance information provision existing and potential passengers would find useful. It would be interesting to hear any insights from our industry contributors.

Do you mean things like ...

Bristol Temple Meads to Keynsham is timetabled to take 7 minutes.   Keynsham to Bristol Temple Meads often takes 11 in the public timetable.

Portsmouth Harbour to Portsmouth and Southsea takes 4 minutes.   Portsmouth and Southsea to Portsmouth Harbour takes 8.

Probably more like dwelling at station X in case there were delays elsewhere. They started doing that at Reading, before all the works began, as you'd often sit outside for ages, making all subsequent stops late. Didcot Parkway is another one (and is the main reason my other half drives, rather than getting the train and "sitting there doing sod all for 10 minutes" :))