Holiday firm TUI has implemented Snowflake as a key element of its three-pronged digital transformation strategy and is reaping the rewards from a consolidated data platform.
That’s a big break from the not-so-distant past, notes Marc Jennings, CIO of Analytics and AI at TUI, who says the company was previously reliant on disparate technology systems. A digital transformation strategy was implemented in 2017 with the aim of consolidating separate systems across the group into a single source of truth for data, as well as eliminating legacy solutions, cutting costs and creating a digital platform for long-term business change.
AWS was introduced as the company’s cloud platform in 2018, followed by Snowflake for data consolidation in 2019, and Tableau for Business Intelligence in 2021. Snowflake’s platform sits at the heart of this program and has helped shift the focus of the work his team undertakes and the benefits they create, according to Jennings:
There have been challenges in the past with delivering on time, delivering to cost, delivering into value and meeting our service levels. The use of Snowflake, and the ability to go fast, build a team and test stuff, is key. We’ve had problems in the past being able to do full performance and operational testing and to understand what that’s going to cost. The ability to now do that has given the business confidence that we can deliver on time with the capabilities and quality that’s expected.
TUI’s core Snowflake product right now is the Data Cloud, intended to create a cross-cloud platform to eliminate data silos. Jennings says this platform gives the business access to a set of workable and scalable services that it can use to respond proactively to fast-changing macro-economic conditions. A combination of Snowflake and the cloud means cost calculations that used to take days now takes minutes:
In our old world, we could only run a process once a month, and it would take a minimum of 24 hours and sometimes it would take days to complete. We were just unable to deal with the size and scope of the volume of changes that we needed to make. Now, we can run that process much more frequently, such as a number of times a week. So, our costs and our prices are much more relevant and accurate.
As well as its exploitation of Data Cloud, TUI is also exploring Unistore, which is a new workload approach in Snowflake that allows organizations to work with transactional and analytical data on a single platform. As discussed at diginomica before, Unistore is intended to extend the applicability of the Data Cloud and make it easier for organizations to share and enrich the information they hold. Jennings says TUI is currently exploring Unistore via private preview:
We’re very excited about what that product offers. It's not just about processing large numbers of transactions and billions of rows of tables that we update, we also have users who have to do single transaction updates. We see that technology as being part of our structure and strategy as we move forwards.
When it comes to other functionality, Jennings says his team is not currently tapping into Snowflake Marketplace, which is a collaboration space that allows companies to exploit a variety of third-party datasets. However, he believes there is potential for TUI to explore this development in the future and to think about how external sources might be used:
There's a lot of data out in the world. As the Marketplace grows, we'll be looking to potentially consume external data that we want to use within our systems. There’s lots of opportunities for us to use other Snowflake tools in the future.
Enabling smarter decision-making processes
Jennings’ team ran a number of proof of concepts around the time the firm selected Snowflake in 2018. The platform was chosen because of its tight correlation with TUI’s data-led business requirements, he explains:
You have to understand the problem or the opportunity that you are trying to solve. Never pick a product just because it's a good product. Understand the opportunity first and only then go and select your product. That was our approach and we understood what we wanted to do. Without set ambitions, goals, visions and capabilities, you’ll buy a product and you won't use it.
Jennings says the adoption of Snowflake, and its use as part of a digital transformation platform alongside AWS and Tableau, means that TUI now has a consolidated view of its data that can be used to enable smarter decision-making processes:
Take abandonment rates on the website. We might want to think about, ‘Why did that happen?’ You need to look at the different customer segments, booking profiles, and the trends – being able to consolidate that data, and then deliver content to the people who need it most, is critical to success.
Another example of the important role of data, says Jennings, is maintenance schedules. TUI has to think carefully about how it checks its fleet of aircraft on the ground. Any timing issues can mean its planes spend less time in the air, which affects the services it can offer to customers, with a potential impact on the bottom line. Timely insight means TUI can create an automated approach to maintenance schedules, says Jennings:
All of this data lends itself to us operating much more effectively and means we’re delivering the data and the insights that our internal customers need to operate the business. And then, ultimately, we can offer better products, better services and optimized prices to our customers.
Jennings advises other digital and business leaders who are thinking of adopting Snowflake to focus on building strong internal expertise. One of the things his organisation has suffered from in the past is an assumption that a well-sold technology product will provide an out-of-the-box solution to business challenges. Jennings says managers buying technology must be careful to avoid a similar fate:
You need to build standards and processes, so that people who've got insights can share their knowledge of how the product is used. Create communities of practice – we've done that now, so all our Snowflake developers and Tableau users can all input their knowledge, they can share their ideas and their problems, and that means we can really make sure we maximise what we do with the product.