Chad McDermott - stock.adobe.com

CIO interview: Marc Jennings, CIO of analytics and AI, TUI

The international travel group is overhauling its data stack and aims to provide self-service analytics to key employees across the organisation

This article can also be found in the Premium Editorial Download: Computer Weekly: How to protect against ransomware attacks

During his two-decade stint with holiday firm TUI, Marc Jennings, CIO of analytics and AI, has seen the company undergo rapid transformation. Through a series of mergers and acquisitions, the German-headquartered firm has grown into one of the world’s largest travel and tourism businesses.

The pace of change has continued to increase over the past two years, as technology at TUI has transitioned in line with the wider organisation and its shift from a firm that is based around local and regional systems and departments to a global IT organisation. Jennings’ group-wide role covers all the analytics and artificial intelligence systems, platforms and processes within the group.

“The company has evolved dramatically over the last 20 years,” he says. “Throughout that journey, I’ve been lucky enough to work with some fantastic leaders who have guided me along the way. I have moved roles roughly every couple of years, so I’ve had new insights regularly and picked up lots of experience.”

Jennings joined TUI as an extract, transform, load developer. He progressed through a series of technical, management and director-level positions before assuming his current position in October 2021. He reports into TUI’s group CIO, Frank Rosenberger.

One of the big priorities during the past year, says Jennings, has been rolling out some key applications that provide the global business with pricing, yielding and consolidated data held in Snowflake’s cloud-based technology platform. This insight is used to provide key performance indicators (KPIs) to the board and information to the financial markets.

Jennings says this data-led transformation is a big shift. The company is moving from a situation where individual regional markets relied on database and reporting systems running on-premise to a consolidated approach that focuses on three core technologies. As well as Snowflake, TUI’s data stack includes Amazon Web Services (AWS) for cloud and Tableau for business reporting.

“Our global transformation has begun and we have started delivering reporting solutions on time, to budget and to agreed service-level agreements,” he says. “Building the technology stack that we’ve chosen for the future of the business is certainly one of the key deliverables and successes that my team has had over the past year.”

Beyond that transformational work, Jennings says his team has also “reinvigorated” TUI’s data strategy, which sets out the data-led objectives for the organisation through to 2025. “We are working through that and it’s very much a journey,” he says.

“But you have to start somewhere and we’re very keen to get that initiative moving forwards. Our goal is to deliver truly data-driven decisions for the whole business. We have a plan and a roadmap to go and execute on that aim during the next three years.”

Selecting the right platforms

Jennings says there are multiple reasons why the business chose to base its data strategy around Snowflake, AWS and Tableau. He says it’s easy to scale these cloud-based technologies, they can be used in combination to support high-performing business processes, and they are easier to support, leading to an impact on cost management.

Another key benefit of the company’s shift to a three-pronged data stack is capability, he says – it can be tough to find workers with expert skills in older systems: “We have had challenges recruiting people for some of the legacy software that we used to rely on in terms of being able to use the software and go faster.”

While TUI has chosen to focus its data stack on three core technologies, Jennings says it is important to state that it is not a closed shop. If other suitable tools enter the market, his team will consider their potential benefits. However, the aim of focusing on a core range of products is to build a stable platform for change.

“They are there for the long term,” he says. “We know the that cost of migrating off a solution or having a product that you don’t use very well can be horrendous. I need to continue building a team of people who can support and develop applications, and who are getting domain knowledge around the products and the tools they use.

“In the past, I’ve had people working all over the place, and we ended up with single points of failure where people move on and you can’t backfill that position. The tool stack I’ve created over the last 18 months has really enabled me to build a team that is fit for the future. It’s also a set of partners that work very well together – and that also really helps.”

Supporting smarter decision-making processes

As well as his day-to-day responsibilities, Jennings says his priority during the next 12 months is to hone the new capabilities that his team are building across their strategic data stack.

“That’s all about delivering great decisions to the business,” he says. “We have a statement internally that says, ‘no weasel words’. We try to avoid – and I constantly hammer this home to my team – just saying things are slower or faster, and instead to ensure that they always provide context for the change.

“So, that contextual information might be a percentage increase or decrease and it will relate to a KPI. We try and do this internally in my team, and then we are trying to educate the rest of the business to be able to do this as well. But to do this effectively, you need data.”

Read more data leader interviews

By being able to make the most of Tableau and other specialist data technologies, Jennings hopes his team can create a set of easy-to-use interfaces that enable other parts of the business to start doing their own data analysis work on a self-service basis.

“That’s about asking, ‘How do we make this work easier for them, so that they can go and do things for themselves?’ We don’t want to be a bottleneck and neither should the technology be a bottleneck,” he says.

“Our goal is very much to work through and understand the tools, the techniques and the processes to establish self-service analysis in the business. Filling that gap between either end of the spectrum is our challenge. We want people at TUI to be able to be informed in the decisions they make.”

Keeping costs under control

Jennings says the result of the digital transformation programme so far is that line-of-business employees now have access to a range of useful services that provide a wealth of information. He says moving from disparate and localised systems to a consolidated date lake via the Snowflake Data Cloud means the business can make more informed decisions.

“In the old world, when we were looking to update data and prices on the website, it used to take two hours,” he says. “If a booking was made, and we wanted to reflect the impact of that booking and make any price changes on the website, it would take in excess of two hours. Now it takes less than 20 minutes.”

Jennings recognises that creating an integrated data stack does bring its challenges. IT leaders who are not careful will see compute demands spiral out of control very quickly. At TUI, data-led jobs can run for hours, so Jennings uses a dual approach that focuses on communication and forecasts to keep costs under control.

“Everything we do needs to be visible and communicated,” he says. “So, if you’re running a performance test, make sure its impact is assessed and communicated, so people are watching how it takes place. That’s about providing the manual oversight, so everyone is aware of what is happening.

“We also have built-in automated forecasts. So, we have modelled out our forecasts, we’ve set alerts and targets, we’ve built dashboards and reports, and we do that across all of our environments. This process means we know how much we should be spending in development. That has clearly taken a while to build up, but it’s critical to our success.”

The key message for other IT leaders who are looking to keep costs in check is to ensure the business knows what data processes it is running through a series of manual and automated checks, says Jennings: “Understand what you’re doing, impact-assess it, set a target, measure it and optimise – that’s our approach.”

Adding tactical solutions to business challenges

Jennings looks back on the work his team has achieved in the past few years and says that despite rapid progress, there is still more work to do. While the technological foundations are in place, he expects his team to source other technological solutions to the business’s data challenges.

As well as Snowflake, AWS and Tableau, he expects to deploy other cloud-based services when the time is right. “We know they are not going to be the only tools we use as we build more sophisticated applications and insights,” he says.

One of the key things that Jennings’ team has learned, however, is that any technology purchase must be accompanied by a thorough evaluation process. He says one of the things his company has suffered from in the past is a preconceived belief that off-the-shelf products will be implemented and used easily and effectively. This is a common problem for other IT leaders and Jennings says it is important to avoid making hasty procurement decisions.

“My advice is, don’t buy too many systems,” he says. “A salesman will come over and say, ‘This is going to be the latest and greatest thing’, but it often solves such a small challenge. And over time, your ability to be able to use it and support it effectively diminishes, and then you’ve got an overhead around your neck that you’ve got to carry.”

The aim, says Jennings, will be to find technologies that complement the existing stack, which generate insight and allow TUI’s employees to feel more confident about the decisions they make. He adds: “The ability for us to source and consolidate information into one place gives us the ability to deliver new insights that we didn’t have before.”

Read more on Big data analytics