Case study

Insights into collaboration with Greater Wellington Regional Council

How Greater Wellington Regional Council used data transparency to build trust with operators and improve network performance.

3 min read

June 11, 2024

Greater Wellington Regional Council’s (GWRC) transit network, Metlink, connects the wider Wellington region in New Zealand. Today, there are more than 38 million journeys a year on the Metlink bus, rail, and harbour ferry transport network.

Challenges

Much like the Bus Services Act changes that enable franchising in the UK, in 2013 New Zealand created the Public Transport Operating Model (now replaced with the Sustainable Public Transport Framework), which empowered authorities like GWRC to take control of designing networks and then tender for their delivery.

In 2018, GWRC implemented significant changes to the transportation network. These changes aimed to enhance network capacity and involved modifications to operators, contract structures, routes, and fare policies. However, the implementation process was difficult, and the network's performance faced heightened scrutiny.

What was previously a highly trusted system fell short of high customer expectations, and the change was extensively covered in national media. Tensions between operators and the authority made it difficult to find a constructive solution. Without an agreed-upon single source of truth, deciding on a path to improve the network was a challenge.

To restore passenger trust, GWRC needed to get their team on the same page as operators and solve the immediate problems in the network. Looking ahead, they needed to enable contract management and continuous improvement of the network.

How we helped

Snapper Services is a trusted technology partner to GWRC, providing ticketing as a service since 2018. Following the network changes, Snapper Services saw a 400% increase in customers requesting refunds for incorrect fares and penalties, and issues with new routes. Snapper Services and GWRC recognised the need for a solution to investigate this large increase in customer queries.

The goal of this solution was to deliver reliable and objective reporting for all participants in the delivery of the network. This was seen as crucial for rebuilding trust between GWRC, operators, and the public. Enabling a collaborative approach would enhance network efficiency and help rebuild public confidence in the service.

Snapper Services developed an analytics service (the precursor to Mosaiq Insights) to help GWRC. The tool enabled operators to have secure access to their own data, so they could collaborate effectively with GWRC to monitor and improve performance.

Snapper Services and GWRC had access to a wide variety of data feeds including scheduled and actual trip performance and factors that could impact the driver’s ability to meet KPIs. The system processed dynamic data (from RTI and ticketing systems) and static data (PTO planning and service design systems) along with third-party weather, traffic and social media feeds. This data was interpreted by both algorithms and experienced analysts to understand if the right bus was in the right place at the right time for every trip.

Results

Establishing trust in data was the first step. The investigation into data reliability revealed that the equipment was sufficiently accurate. However, there was a pattern of incorrect usage, including incorrect trip information and nonconformance to routes. While driver equipment training was the priority, the granular data made it clear to all parties that the issues were not deliberate. Further analysis showed that drivers, under stress, were trying to correct their errors and issues during the trip. Understanding this enabled a productive and empathetic conversation with operators about driver training and capability.

From here, operator contracts had agreed-upon KPIs that needed to be applied to trips. As equipment usage improved, GWRC and operators were all able to turn their attention to where KPIs were being missed. With a neutral, trusted, and clear understanding of performance, GWRC and operators were able to hold constructive conversations about how improvements could be made. This opened the door for the continuous improvement process that all parties were aiming for, and that underpins the relationship today, which is based on trust and transparency.

Crucially, the performance of the network improved substantially across a range of metrics as a result of the service being consistently used by GWRC and the operators: punctuality rose from 87% to 94.8%, and reliability increased from 94% to 98%. Most importantly, passengers’ confidence in the service increased from 85% to 90%, a strong result following a period of deeply felt disruption.

The technical side: building and deploying on AWS

AWS Architecture

From an architecture perspective, the Insights platform is deployed using the appropriate AWS services.

Data feeds are ingested through API Gateway endpoints, which are backed by AWS Fargate for serverless compute which can scale when needed. Ingested data is stored in DynamoDB where the data is short-lived and processed in the near future, and in S3 where the data needs to be persisted for longer periods of time.

Data processing is undertaken using Lambda functions orchestrated by Step Functions that run on a regular schedule. The Lambda functions can scale out to meet the amount of data that needs to be processed in each run. The processed data is stored in S3 and RDS, ready for display to the user.

The user interface is delivered through a web portal deployed on Elastic Beanstalk, which can scale to meet user demand, and provides the functionality that supports the user workflow as analysts across the organisations review and update bus trip details relevant to KPI calculations.

TCO Analysis

As an existing SaaS product, no TCO Analysis was undertaken to deploy the analytics tool into GWRC’s system.

Since these services were developed and deployed on the AWS cloud, GWRC does not need to look after internal infrastructure, and they enjoy unlimited user seats.  The services leverage the Mosaiq platform to ensure that ticketing data, vehicle position data and capacity data is processed as efficiently as possible.

What we learned

Data volumes can vary tremendously from one SaaS customer to the next. Having an architecture that can scale easily is paramount. 
In the context of evolving contract management and franchising around the world, being able to manage data access in smart ways can reduce waste, cost, and increase the speed to being able to access data for public transit authorities and operators. 
Automating not just the application itself but also the DevOps monitoring stack from the beginning saves time and increases trust from an operational perspective.
There is no one-size-fits-all at any tech stack level. Always seek the right tool for the job. Not everything needs to be a container, and not everything can be reduced to a few lambdas. Same for persistence: always start with the customer need.
With an organisation of this size and complexity, it was important to engage with the end customer early to encourage collaboration and open comms channels.