A Day in the Life of an AccorHotels Data Scientist


What is Data Science? According to the UC Berkeley School of Information, the field is emerging in the “intersection of the fields of social sciences and statistics, information and computer science and design.”

In the travel world, the rise of cheap and ubiquitous data has shown a growing shortage of professionals in the field. A McKinsey Global Institute study estimates that there will be four to five million jobs in the United States requiring data analysis skills by 2018. And the executives need to be able to ask the right questions and understand the results of big data analysis effectively. Also, more and more tourism brands recognize it and, at a recent EyeforTravel event in the United States, 67% of participants interviewed said they believed that investing in analytics and data would help them retain customer loyalty – the driving force behind their distribution strategy!

Science and data analysis are being taken seriously by AccorHotels. But what exactly does a data scientist do? Kevin Tran-Dai, Accor’s chief data scientist, says that a typical day involves:

  • Meeting with internal customers to identify needs or challenges, which can be:
  1. Increasing customer loyalty
  2. Measuring advanced metrics such as incremental sales for campaigns or special offers

III. Analyze customer feedback to detect negative points during stay

  1. Identify travel trends using open data to increase the efficiency of search engine advertising by country of origin
  2. Develop a 360 ° view of a client to detect triggers for dynamic pricing or deduplicate (the practice of eliminating repeated copies of data) partner databases
  • Discuss new models and tools with other data scientists or business analysts and to define which methods and data to use to build analytical tools or study.
  • Improve code and deal with bugs and technical issues

“More than 70% of my time is used to merge data, deal with missing values, sanitize them, develop and test models.”

One area where Tran-Dai’s energy is focused is advanced analysis. This, he says, is being used in three main ways to:

Understand and predict customer behavior or preferences to improve your experience and make smart recommendations. ‘Next stay score’ is one way to determine the likelihood of customer booking a stay in the future by using historical data and weak signal data. This could include stay sequences and online search patterns. Customer lifetime value is also determined by using data to predict possible next destinations.

  • Test and build pricing optimisation algorithms that challenge revenue management systems. In doing, so the aim is to help revenue managers maximise hotel revenue. Areas of focus include: demand forecast, elasticity estimation and lead-time triggers.
  • Gather or analyse external data that has a direct impact on performance such as events, exchange rates, weather, tweets or reviews. Tools used include: EventFul, OpenweatherMap, QuandDL, TripAdvisor, Tweeter.

What is clear is that the hotel landscape is changing rapidly and Accor recognises that in order to survive into the future, as a group selling services and brands to hotel owners, it will need to up its game. “As a service provider, our main role is to provide an efficient environment that allows hoteliers to focus only on their business and customers during the stay. Data is certainly a way to optimise but it relies on a strong operational base and systems,” stresses Tran-Dai.

So what is the role of hotel groups like Accor going forward? In Tran-Dai’s view they must:

  • Deliver strong IT systems and software to hotels (easy deployment, complete integration, best in class features and so on).
  • Provide and share a 360° view of customers to each hotel to allow personalised services.
  • Define and share best practices related to, for example, operational training, processes and so on.
  • Manage or provide distribution channels such as web, OTA contracts; the aim here is to find the right balance between all the available channels out there that sell rooms and services. Yes, source partners for extra services but also have your own system/channel to sell on.
  • Have a strong brand with a clear ‘service level’ promise.

Step by step

According to Tran-Dai, the steps Accor is taking to date are working well, and data science really can and does reveal value. However, it’s not all plain sailing.

“The main challenge is to have a sufficient quality of internal data to deploy models and insights and to have this running in our systems such as the PMS, web and so on. But if the company is not mature enough to provide accurate internal data (such as customer information, bookings, stays) there will be too much input noise that will strongly decrease the accuracy of any prediction or recommendation,” he says.

Another concern is whether the traditional hotels are truly aligned with the needs of the new generation. However, Tran-Dai believes that the luxury sector will continue to grow, as it is here that hotels are still able to offer a truly differentiated service and experience.

Going forward, there will be a lot of opportunities as the market is changing rapidly. However, the hotel companies that succeed will be those that are able to deliver excellence in operations, truly understand expectations and deliver those at the right price, concludes Tran-Dai.

Source: Eye For Travel

This site uses Akismet to reduce spam. Learn how your comment data is processed.