Case study: Manchester Airports Group using AWS to revamp its data strategy

Hoovering data

Initially, MAG took a bottom-up, technology-led approach, hoovering up as much data as possible, according to MAG chief digital officer, Ryan Cant.

Once the foundational aspect of the technology was in place, ensuring there was one data lake and one visualisation tool rather than multiple versions, the project needed to pivot to top-down.

This allowed the team to think about different functions of the business like operations, commercial or people, and their desired business outcomes that required better data insights.

We can ask for all the data in the world, but if the application isn’t capable of generating it because we’re not capturing it properly, it’s not automated, it’s not got the right API connectivity to get the data out – then we’re on a hiding to nothing

The company assigned product teams aligned to application stacks in support of each particular business area. These teams understand the specific people, processes and customer outcomes for that function, what it means for partners and airlines, and how the technology enables that.

“And importantly, what data elements are required to best optimise that area of our business and how do they come out of the application stack that we’ve got,” Cant says

“We have gone with a fairly integrated strategy in terms of each product area [being] accountable for not just its application stack, the roadmap around that, and the business functions that it supports, but the data outcomes as well – supported  by an underpinning data technology capability.”

While this has left MAG with separate product teams, the company has avoided data silos as a result of having a single data technology team pulling it all together. But ultimately, the vertical product teams are empowered to define what data outcomes they are trying to achieve and why, and how their applications need to enable that.

“Because at the end of the day, the data will only come out of applications,” Cant says. “We can ask for all the data in the world, but if the application isn’t capable of generating it because we’re not capturing it properly, it’s not automated, it’s not got the right API connectivity to go and get the data out – then we’re on a hiding to nothing.”

Data consistency

All this data goes into a centralised data lake, managed by a team responsible for not just the technology side, but also the data structure element, to ensure data consistency. If a passenger in one product area is defined in a certain way, that definition has to be consistent everywhere across all MAG airports and applications.

While this approach means MAG can avoid data silos, it is facing a challenge with the various applications it has yet to integrate or transform.

Over the next 12 to 18 months, the organisation will pick off the important applications and ensure they are integrated into the data stack.

One of the potential targets is MAG’s asset management system. The airport group is ultimately an infrastructure operations business, and is very asset-heavy.

The firm, which uses IBM Maximo for all its asset management, is keen to ensure that it integrates the right data points from that into the lake thoughtfully, rather than simply taking all the data in Maximo and attempting to make sense of it afterwards.

MAG is using an AWS data lake on an S3 environment, with Redshift on top of that for business logic. The Redshift streaming product allows the company to bring both batch and real-time data insights and analytics together, giving the reassurance that the same business logic is being applied to the data after the event and in the moment.

“If you are an operator looking at a live view of what’s happening now, you’re getting the same trusted view as you’ll get on the report after. That sounds really simple, but it was quite complex to achieve that one, both from a technical perspective and a business data logic,” Cant says.

Getting into generative AI

The organisation is trying out Bedrock, Amazon’s tool for building generative artificial intelligence (AI) applications, to enhance the passenger experience. The aim is for a prototype customer service agent powered by natural language to provide intuitive interfaces for passengers to help them make their way around airports.

MAG is also working with AWS on innovation around computer vision AI, which takes unstructured data and images, and turns that into structured data. The company is using the technology to see whether it can start to capture events happening out on the airfield – for example, a plane on a stand.

Currently, MAG is reliant on a complex ecosystem of airlines, ground handlers and airport staff to get this information. Operators – who may or may not work for the airports – are required to fill in semi-manual, semi-automated processes so MAG can understand the latest status on the turnaround journey of an aircraft. The opportunity to get a trusted data stream for that process would be incredibly helpful to the organisation.

Another proof of concept is being tested for monitoring baggage loaded on and off planes, which will let MAG predict when an aircraft will be ready to depart – and provide real-time information about waiting times in baggage recovery halls.

“Some of the computer vision models and the ability to go and very quickly develop new models is one of the things where cloud computing gives us a really interesting opportunity,” says Cant. “The concept of computer vision has been around for quite a while, and that use case in airports has been around for quite a while. Where we see the opportunity with Amazon is being able to get there quickly and to a very high quality.”

Overall, the aim of these innovations and experiments are technology building blocks towards a great passenger experience and a great colleague experience, something MAG sees as two sides of the same coin.

“Motivated, happy colleagues are the folks who are going to make your journey through our airports feel good. The tech alone will not make your journey good, it’s the tech enabling our colleagues to deliver excellent service,” saysCant. “As we join up some of these technologies, the opportunity to delight passengers is pretty good.”

Passengers should experience a good flow through MAG’s airports, with all the parties in the airport geared up to process them as efficiently as possible.

“That’s a foundational piece and I’d like to think we’ve got that right over the past year or so in a way that’s quite obvious now to our passengers,” he adds.

Cant pointed toward some of the organisation’s Net Promoter Score ratings, which he said place Manchester Airport Terminal 2 higher than Apple, as a proof point of the success of the organisation’s efforts to modernise. 

However, Manchester Terminal 1 and Terminal 3 both fare badly according to the latest Which? customer survey of the UK best and worst airports, showing the importance of testing out tools like Amazon Bedrock and computer vision AI to enhance the passenger experience.

Cant notes that, according to MAG’s own data, all passengers across Manchester and Stansted airports have been processed through security in under 15 minutes since October 2023.

“That’s as a deliberate focus on technology investment alongside operations colleagues, who’ve been [focusing] on delivering that level of service. It is absolutely about delivering the basics, and then having won the right to play by enabling those basics, then delivering real added value experiences,” he says.

Personalising traveller needs

One enhanced experience MAG is hoping to offer is an environment more personalised to traveller needs and the reasons for their travel.

In other consumer worlds, people can generally be segmented into a certain profile and then organisations can tailor a service around that – from a travel or passenger perspective, this is more of a challenge, Cant explains. When people travel for business, they have a different persona to when they’re travelling with family and friends.

“It’s being able to use technology and data to start to understand the different profiles that are passing through an airport on a particular day, and then being able to tailor the services that we offer and promote to you, [such as the] mix of food and beverage offerings that our partners provide. That is increasingly data-driven and to a certain extent can become a bit more dynamic,” he says.  

The airport group has also teamed up with AWS and Ryanair to develop a near-real-time event platform, aimed at improving the passenger experience at London Stansted and Manchester.

The cloud platform drives information updates out from core airport systems, so passengers receive automated gate announcements via departure boards and screens in the terminal at the same time they get the update on the Ryanair app.

The intention is to expand this technology to other airlines in the future. The group currently serves more than 250 destinations, with almost 60 airlines flying from Manchester alone.

“The more we can collaborate and share using cloud-based technology as a foundation for that, the better for all our passengers,” Cant says. “Ultimately, the passenger expects the airline and the airport to work collaboratively together in service of delivering them a great experience.”