Open-source journey planning - the way forward?

Peter Bell, managing director of journey planning provider Trapeze Group, ponders the business models which will underpin future travel information services from a UK perspective

Traditionally, journey planning websites for public transport in the UK (for example, Transport Direct, the Traveline regions or National Rail Enquiries) have been provided by the transport operators keen to increase ridership and revenues, or by public bodies who hope to encourage a modal switch to public transport by making it easier for users to plan their journeys on public transport.

However, in recent months the idea has been gaining pace of simply releasing the data and allowing different suppliers (including one-man app developers) to implement applications with the data. People have been talking about this as an 'open source' model for journey planning. There are advantages to such an approach but there are also issues which still need to be resolved.

Background

Today, journey planning provision for public transport is outsourced to a small number of specialist providers. The service usually encompasses collating the data from a variety of sources, identifying and fixing any data abnormalities, developing and maintaining the algorithms to plan the journey, and the provision of the actual website itself.

Arguably the change to open-source journey planning all started with former UK Prime Minister Gordon Brown's Open Data Initiative. Launched in October 2009, the aim of the initiative is to make government data available to third parties free of charge for them to use and exploit as they see fit. To date, though, only a limited form of the timetable data has been released on the Open Data website.

1466 Transport for London (TfL) announced its intention to release significant parts of its transport data in June 2010. To date, the data made available is not suitable for journey planning but TfL has indicated its intention to also publish its timetable information. Greater Manchester Passenger Transport Executive (GMPTE) has already gone one step further and is making its bus timetable data available in a universal format known as ATCO-CIF.

By contrast, National Rail Enquiries has faced significant criticism for cutting off independent iPhone App developers from a so-called free feed that these developers were using to provide alternative tools for checking the status of rail journeys; the general media certainly believe that independent iPhone app developers will produce more innovative apps in shorter time with more features than large organisation such as National Rail Enquiries.

How it might work

With all this going on, it is not surprising that some people are now floating the idea of 'open source' journey planning. The idea is simple: operators release their data to the community of developers. Developers then independently develop leading-edge applications.

Supporters believe that such an approach should lead to a more vibrant competitive environment and, therefore, better offerings to the travelling public. The independent developers will be competing in two ways in the market for providing journey planning services. Firstly, they will be competing with each other over the quality of their applications - particularly the usability and flexibility of applications. Secondly, they will be competing to develop novel business models. So some developers may choose to simply sell smart phone apps for a fixed fee. Others, particularly those offering services through websites, may choose to offer location-based services around the journey planning. So the user plans a journey, and the website offers to book hotels and taxis as part of the journey plan. A good example of this is the Trainline app for iPhone, which provides the ability to plan a journey and then book the journey immediately on the iPhone.

However, it is the chance for new business models to develop that is the greatest attraction of the new approach. A diversity of business models will drive a range of different services allowing more tailored services for the travelling public.

Challenges

While this approach certainly has some attractions, experience to date of providing journey planning systems would indicate some issues that need to be resolved.

The first issue is the interaction between journey planning algorithm and data. These days it is rare that a journey planning engine from one of the established specialist providers generates an unreasonable journey plan, unless the planning data has some error in it. Interestingly, different algorithms will often expose different types of errors. When data curation and journey planning algorithm development are part of the same operation this type of issue occurs infrequently and is fairly easily resolved. However, in the brave new world of open source timetables the curators of the source data may suddenly find themselves dealing with 10, 15 or even 100 developers contacting them with questions around data quality. Responding to these queries in a useful manner will require different business processes to those used today and will likely also require additional resource to deal with the increased queries.

Secondly, how is the quality of the planned journeys assured? Imagine the furore if, for instance, somebody misses an important hospital appointment because the interconnection between two services is not properly taken account. This is more of an issue than might appear at first. Although simple at a high level, journey planning has numerous subtleties that need to be taken into account of. For instance, taking correct account of changing bus timetables, particularly around Christmas, requires careful handling. It is easy to see a new entrant launching an application that simply does not get these subtleties right.

A related issue for journey planning websites is ensuring that they provide a reliable service 24 hours a day, seven days a week, whatever the circumstances. One of the challenges faced by providers of journey planning websites is the variation in usage of such sites. Extreme weather events drive usage of the sites up exponentially, as do other events such as large sporting events. A potential scenario in this new world of open source journey planning is that established providers are replaced with agile, innovative companies who have invested heavily in the look and feel of service provision but less in the backbone infrastructure. Nobody would realise the issue until the day an abnormal weather event hit and the service's users suddenly found themselves without any journey planning capability.

In the early days of computerised airline reservation systems these were run by the major airlines. Smaller airlines would then buy into a system. One of the major airlines was accused of giving itself an unfair advantage by listing possible journeys using its route ahead of better journeys (from the customers' perspective) which used routes from other airlines. While prioritisation of routes is not intrinsically wrong - after all, we all use Google where the sponsored links are at the top of the page - this needs to be disclosed to the traveller.

Other concerns focus around potential market failures.

The first issue here is ensuring social inclusion. There is a significant potential market for public transport journey planning in London. Likewise, the other major conurbations in the UK can probably support independent journey planning. By contrast, it is unlikely that a commercial journey planner will be able to generate much revenue from rural areas elsewhere.

The second issue arises out of avoiding a commercial monopoly developing. One possible way this may play out is that lots of small independents move in, destroy the market for the current incumbents, ride the wave of initial interest, then lose interest and move on to new areas leaving a vacuum and a frustrated public.

Possible ways forward

Clearly, this new world offers many exciting possibilities and 629 Trapeze is among those companies looking at how to exploit new business models. Nonetheless, without some careful planning the public could lose out.

To avoid this, the government should consider two options. Firstly, in the same way that basic Internet access is fast becoming a human right, so access to a basic journey planning service (either by web or phone) should be guaranteed. This service will probably need to be financed either via a subscription model, or via some form of government grant. Such a basic journey planning service would also set a minimum standard that would encourage innovation to move beyond it.

Secondly, for the newer services that exploit new business models, some form of economic regulation may be necessary. The obvious option is that access to the journey planning data is licensed. As part of the licence the regulator might, for instance, publish test data sets with the journey planning data. Developers of journey planning software will then need to show the regulator that their algorithm works reliably on this larger test data set. Only if they have proven their algorithms to be reliable will they be licensed to provide a journey planner to the public.

Whether such economic regulation would be acceptable to the supporters of open source journey planning remains an interesting question.