Rail Professional Magazine Editor Sam Sherwood-Hale spoke to Alex Froom about personalisation in the transport industry and how rail should be using data to improve its service…
What is Zipabout and why was it created?
Zipabout grew out of a project we built for TfL and the London 2012 Olympics. Back then I was running a creative agency working in digital engagement for consumer brands and Dan was involved in developing complex data platforms for the online advertising industry and companies such as Thomas Cook.
We’d worked together on a series of customer service initiatives for TfL when they asked us to create a demand management platform for the Games – we had six weeks to dismantle the standard journey planner approach of ‘quickest route first’ and balance it with predicted usage and capacity, hour-by-hour through the Olympics.
It was only after the dust had settled that we realised nobody else seemed capable of using high volumes of disparate data to create personalisation and capacity optimisation for the transport network! Around about the same time, we also realised that we personally needed an app that could tell us if we had time for another pint before catching a train, which is basically personalised journey planning.
We thought there was something in it, so we started Zipabout. The result is a communication platform that processes every piece of data relevant to the transport network – from the obvious operational inputs, such as live vehicle locations and weather, down to the human elements like behavioural patterns and sentiment analysis.
What is Zipabout doing to solve the issue of personalisation in the transport industry, how does it work and what makes it unique?
Coming into the transport sector was a bit of a surprise – I think we had made a lot of naïve assumptions around technical capabilities and digital expertise. Running trains safely around the country is far more complicated than anything we had ever done, so it was a surprise to find an industry that appeared to be 15-20 years behind the digital consumer sector.
I think that was probably an advantage though, as it’s allowing the more forward-thinking rail companies to ‘jump’ into the state of the art without legacy.
Our approach to personalisation comes from the online advertising industry, where behavioural understanding and personalisation have been the focus for over 15 years, and from B2C engagement where brands live and die by engaging with the intended consumer in the way that suits their lifestyle – be it real or aspirational. Big brands such as Facebook, Google and Amazon have been using ‘big data’ and behavioural understanding to personalise their products for years, and our platform is based on that approach.
The rail industry in the UK is beginning to talk about personalisation, and there are a couple of companies out there that definitely ‘get it’ now, but most think a series of tick-boxes on a website or the ability to save a journey in a journey planner means you have personalised the customer experience. It doesn’t.
What makes the Zipabout technology unique is that it is built from the ground-up for the transport sector. This has led to some great partnerships with companies such as AWS (Amazon Web Services), who have helped us create products that the industry can quickly and easily buy as a service without the usual SME procurement issues or capital investment.
Our platform uses myriad inputs of live data to predict and monitor every element of a transport network – in the UK for example we map live weather and air quality data to events, live running train services, bus performance, traffic congestion, Facebook data… we process every tweet in the UK and map the relevant ones to transport services, and some of our services are supporting 40-50 interactions a second. We have 4-5 years of it now, so it’s a lot of data!
What is true customer personalisation?
I’m not sure that ‘True Personalisation’ is actually a recognised phrase or just something we have coined to try and explain the difference. The best example is something like Facebook or Google adverts – you see content that is relevant to you and aimed at you based on the interactions and decisions you have made. Some of these are obvious and easy – if you ‘like’ a certain brand on Facebook for example, then assumptions can be made and cross referenced with your other activities to develop a profile of you that advertisers can access.
Some are less obvious – for example, where you rest your mouse on the screen, which articles you hover over, where you stop scrolling, how you react to a message or suggestion and when and where you look at certain data. This is what our model uses, alongside a healthy combination of behavioural psychology and machine learning, in order to develop ‘true personalisation’.
The reason we make the distinction is the phrase is being heavily overused to describe very simple content filtering. This is not personalisation – it wasn’t even called that in the late ‘90s when e-commerce sites allowed you to save your preferences – and it certainly doesn’t impress the user in this day and age.
What are the ways you see personalisation being utilised in the rail industry?
There are some really quick wins in the rail industry that would make a huge difference to both operators and customers.
For some time now we have had the ability to provide personalised real-time travel updates for the entire UK at an individual level, through any digital medium (Operator apps, Twitter, Messenger etc) and we have developed an API to do just that with support from the DfT. In a nutshell, it allows you to talk to just the people affected by a particular event across any mode, often in advance of it becoming a problem and without adding burden to customer service teams.
By delivering personal updates automatically by ‘push’ not ‘pull’ you can begin to defuse disruption before it even begins and address the information overload that is being pushed through Twitter – perhaps tell thirty people before they leave work that they would be better to take a different train, or a slightly different route, or maybe tell three people on a delayed service to change at the next station and catch the express from platform six.
That’s the easy stuff though – where it becomes really interesting is when you begin to understand the behavioural choices your passengers make and why they make them. That’s enormously powerful, as it means you can not only begin to tackle meaningful behavioural change but also design your services around your passengers without using those dated ‘traveller persona’ models that don’t actually represent anybody.
How is the transport industry, particularly rail, lagging behind other industries when it comes to personalisation, and why does that matter?
There are two issues here, one which needs to be fixed by the industry and one that hopefully we are fixing. The first hurdle is attracting the right people into the industry from relevant sectors – big creative agencies, online advertising specialists, cutting-edge tech firms and UX and UI designers.
There’s a lot of exciting stuff going on in transport these days, but you wouldn’t necessarily know it from the outside, so we need to attract the sort of people who might end up working for Google or Facebook and empower them to make a difference. It still amazes me how rarely you meet anyone in these roles with experience from other sectors.
The second is simply being realistic. Passengers rarely choose an operator out of brand loyalty – there are some routes where that may be possible, but many just choose the train that gets them where they are going aligned to their schedule or budget. Likewise, nobody lives at a station – the rail journey is always part of a more complex door-to-door journey, and that journey is always full of personal decisions and triggers.
Perhaps parking is a problem, or the bus can be late, or the weather influences the traffic, or it’s half term, or there’s a football match, or the toilets are closed after midday. Perhaps the traveller has accessibility issues, visual impairments or gets confused in busy environments. In order to help these people get the best from the railway, you also have to understand every other part of their journey, and that’s not practical for any one operator to try and do.
Do you see technology and ‘intelligence’ as having an inherent issue with personalisation?
Not at all – done well the user shouldn’t even be aware of the process. You could argue that the current broadcast / one-size-fits all approach through social media and comms teams is far less personal than an intelligent technological approach can be. The transport sector – and rail operators in particular – have thousands of customers to talk to, all of whom have slightly different needs, and you can’t expect that to be done face to face all the time.
Done badly, however, unimaginative applications of half-baked tech could be far more damaging than doing nothing. Non-personalised chat-bots that are nothing more than glorified status boards, or voice assistants that don’t understand your personal needs could disengage customers even further as they’re just repeating the same broadcast message. The technology will not replace people though – the human touch is still vital.
If done properly, how can passenger behaviour analytics and two-way communication make the whole rail network run more efficiently?
Ultimately this is not just about personalising communication or customer experience, but personalising the whole railway. Two-way communication means you can really understand each and every passenger and enable them to use the railway in the way that suits them best.
At its simplest level, it enables behavioural change to optimise capacity and increase customer satisfaction – for example, you can encourage users to use less busy services, and even reward them if they choose to.
By fine tuning the way people use the railway – hour by hour, day by day, through disruption to good service – there are huge efficiencies and savings to be made that can save millions of pounds whilst dramatically improving the customer experience.
How is Zipabout already using this technology in collaboration with rail operators?
We have a number of interesting projects on the go at the moment. We have done a lot of work with Transport Focus for example, who were the first to really see the potential in personalisation, and we are currently exploring embedding communication capabilities into National Rail Enquiries and GWR.
Elsewhere we’re about to start a really exciting project with ScotRail, and we’ve also started some analytics work for Govia Thameslink Railway / Gatwick Express and Chiltern Railways.
Looking ahead, what could this all mean for rail operators, and the transport industry as a whole?
I think there is now a really strong appetite to start making a difference in rail – both here and in Europe. A few years ago when we first started demonstrating what we could do there was an attitude of ‘that will never happen in rail’, but that’s changing.
There remain the usual battles with procurement or developments that weren’t in the original franchise bid, but with the work the Innovation Panel has done there is no reason why we can’t create a world-class customer experience to match what is still one of the safest railways in the world.
A specialist in high volume geo-localisation projects and B2C engagement, Froom worked on TDM programmes for the London 2012 Olympics. In 2011, he founded Zipabout with colleague Daniel Chick. For more information or to request a copy of an upcoming white paper by Zipabout discussing personalisation in the transport industry, visit www.zipabout.com