Southern Rail has a UX problem

For the past 3 years I’ve had the dubious pleasure of commuting via Southern Rail.

If you scroll back a year or so in my Twitter feed you’ll notice it’s about 95% rants directed at @SouthernRailUK. I have, however, resigned myself to the fact that nothing I can do is going to change their shoddy service and I’ve largely stopped venting my almost daily dismay at how they can possibly get away with running such an appalling ‘service’.

But this Saturday I was compelled to tweet them, not this time about their train service but about their ticket machines.

If there’s one thing that irritates me more than rail delays it’s bad design and user experience. The first is certainly not my area of expertise, but the second being a reasonably large element of my job it seems perhaps I can help Southern out a little.

So that’s what I’m going to do, and I’m going to record my progress right here, and I’m not going to stop until the problem is fixed dammit!

The Problem

It’s no secret that UK train ticketing is a complex subject. Last month is was announced that a trial will be running from May to simplify the 16 million (yes, MILLION) possible routes and fares and this seems to be dealing with the same problems acknowledged by the government in 2011.

The most recent trial is attempting to tackle the frankly bonkers pricing system that means a single ticket direct from A to C passing through B can be significantly more expensive than two singles, one A to B and one B to C. These are often long-distance journeys and the quirk is known as Split Ticketing

A nation-wide solution to that problem seems years away, but the one I’d like to tackle is more localised and, I think, more achievable.

Watch this video I recorded this weekend whilst trying to purchase a ticket for my wife from Haywards Heath to Victoria and I’ll see you after….

That’s right, the cheapest ticket valid for the journey I was making was on the THIRD page. I very nearly bought the £16.30 one, then noticed the £15.60 one (for Victoria specifically, instead of ‘London’ which I could use for London Bridge as well) but for some reason I decided even that seemed expensive. So I hit the pretty-easy-to-miss ‘More’ button, twice. And there it was the Super Off-Peak Day Return.

Even then there were two different prices and for the life of me I couldn’t work out the difference so just plumped for the cheaper one.

So that’s £8.65 instead of £15.60, a saving of £6.95 or 45%.

I tweeted the video to Southern with the following question:

Hi @SouthernRailUK 1- why is the cheapest ticket on the THIRD page and 2-why are there two super off-peak prices?! NOT customer friendly.

Now I’m used to the spectacularly passive Southern twitter team but their reply was particularly useless.

Gee thanks “^A”, I’ll be sure to send a link to this article to customer services.

Shitty Design, or Dark Patterns?

Now, putting the cheapest option as the last isn’t necessarily a crime of design if it is at the same time the least likely option that a user is going to select. But my instinct tells me that your average ticket machine user selecting ‘London Victoria’ on a Saturday morning is more likely to want to go and come back that day on the cheapest possible ticket than purchase a 7 Day First-Class Season Ticket to all London Stations.

But this got me thinking, is this just pitiful design or are they actively trying to get us to buy the more expensive ticket (whilst presumably being legally obliged to ‘offer’ us the cheapest alternative)?

Dark Patterns are design techniques that intentionally ‘trick’ users into doing something they wouldn’t necessarily do in the general course of things, think leaving newsletter tick boxes ticked, or using double-negatives to get around the fact legislation has made things such as newsletters opt-in (‘tick here if you don’t not want to receive a newsletter’)

How to make it better?

Let’s give Southern and/or their ticket machine vendor (Atos, apparently) the benefit of the doubt and assume this is not a case of Dark Patterns and they’ve innocently missed how sub-optimal their interface is.

Our cheapest ticket is currently at position 22 out of 24 and languishing on Page 3/3. Let’s see how we can use some fairly basic User Experience (UX) principals to improve things, get the total number of pages down and try to make some space on that first page for the cheapest valid ticket.

1-Remove First Class options. I did a bit of Googling to find out some figures for the proportion of 1st to Standard Class tickets sold. It seems the operators aren’t keen to give this out

The Train Operating Companies will not make available such information as they have regarding occupancy levels in first and standard class or the numbers of first and second class tickets sold, citing commercial confidentiality

Let’s be generous and assume we’re talking under 5% of all tickets sold from ticket machines as being First Class. That’s a 95% chance that the person that is making the purchase does not want a First Class ticket so let’s just take them off from the 1-click options. We’ll put a ‘travel in comfort today, upgrade to First Class for £x’ button on the checkout page so the machine can still sell these tickets*

*As an aside this kind of ‘call to action’ might actually increase the number of First Class tickets sold.

So we’ve just removed 10 of the 24 options. Woop!

Target Ticket Position = 13/14 | Target Ticket Page = 2/2 |Tradeoff = People purchasing First Class tickets must make one more click.

2-Remove Redundant options. So this one doesn’t actually get our target ticket any closer to the first page, but it does get rid of two redundant options and reduces confusion.

Remember Q2 from my tweet to Southern?

2-why are there two super off-peak prices?!

It took me a while but I think I figured it out. Having read the text on the video several times it seems there may be two tiers of Super Off-Peak fares, one for weekdays and one for weekends and bank holidays. But the machine knows it’s a Saturday and that we want a ticket for today (you can buy tickets for the next day after about 5pm but this was 10am and in any case that option comes earlier in the ticket buying journey). So let’s lose the Super Off-Peak single and return £9.95/£10.05 options.

Target Ticket Position = 12/12 | Target Ticket Page = 2/2 |Tradeoff = None

3-Reshuffle based on probabilities. So this is how our front page now looks, page two just contains the single and return version of our target cheapest ticket-

At this stage we probably could optimise enough to get down to just one page. We’ve got Off-Peak Day Return, Off-Peak Return and Anytime Day Return for both Victoria specifically and London generally.

The difference between each of those is merely when you’re allowed to make your return journey (respectively the same day, within the next 30 days on an off-Peak train and within the next 30 days on any train). So you might improve the UI by asking the user on this page if they want single or return and in the case of the latter ask them when they intend to make their return journey ‘Definitely Today/Definitely another day Off-Peak/Not Sure”

But let’s come back to that, for now let’s do some pretty basic statistics. What we want to do is sample a statistically significant number of users and base our order on the options they’re most likely to want.

Now I don’t have a statistically significant number of users, I only have me and my 4 month old daughter here right now, but let’s give it a shot. Here are some scenarios we might consider (my instinct in brackets):

  • I’m going to London today and coming back again today (High)
  • I’m going today and coming back again on a weekday outside of the morning peak or on a weekend. (Low)
  • I’m going today and coming back again on a weekday morning some time in the next month. (Very Low)
  • I want to be able to travel on the Gatwick Express (Low, there are no direct Gatwick Express trains from Haywards Heath on a Saturday so this would mean changing trains)
  • I want to buy a 7 Day season ticket (Very Low)

So that High gives a big promotion to our little-old Super Off-Peak Day Return and the tie-breaker of needing the Gatwick Express not being likely puts us in A1 position!

So without doing an actual survey, something Southern could very easily do (or just do a bit of analysis of their own sales records) I think our page one looks like this:

Target Ticket Position = 2/12 | Target Ticket Page = 1/2 |Tradeoff = Users wishing to buy a 7 Day season ticket must click ‘more’ once.

Summary

By following some pretty basic design principals we’ve got the cheapest option at the very top. You might consider removing that £8.60 fare altogether as who cares about 5p but it serves to keep things symmetrical so we’ll go with it.

As previously mentioned, we’re still being a bit unfair on our users here. I’m a fairly well-versed train traveller but your average ticket machine user might not know the validity of those different return fares, so if I were rethinking the whole of the UX I’d probably go for something like:

  • Page 1- Where do you want to go?
  • Page 2- When do you you want to return?
  • Page 3- Extra options (I need Gatwick Express, First Class, Any London etc)

But I think the above is a pretty good start. Now to send that link to customer services…..