The online team, for example, thought of the customer as just one of a number of visitors to the site, while the loyalty department thought of the customer as a registered user. But the company wanted to alter this in order to get a more complete picture of its customer base.
That meant making its email marketing, data warehouse, online and loyalty teams working together on what was initially intended as an IT efficiency project.
"Usually you can't get funding for that kind of project, but we prepared the investment case. We looked at the different use cases of the project with the expected return on investment (ROI), and went to the board and said, 'this is how much we're going to give to the business'," Hotels.com chief technology officer Thierry Bedos told delegates at Computing's Big Data Summit 2015.
But Bedos explained that it wasn't as easy as a "yes" or "no" from the board - as his team had to take a gated approach. "[The board said] why don't you start on the first use-cases and then we'll fund more [if it works]."
The team had 170 potential use cases, and quickly realised that the project was bigger than a user database consolidation. It invested in Hadoop, Cassandra and other big data tools.
"When you put this sort of architecture together it has the potential to boil the ocean. You find yourself with all these different capabilities to deliver to the business, and then you have to think about what is going to get the most money to the business in the shortest time," Bedos said.
The use case it chose was merging data points about customers in order to make better decisions in the online space. The analytics and business intelligence teams were also interested in the project and sat-in on the meetings, which was new for Bedos. "That is the friendly collision between the online and offline world," he said.
Before, the online world would pass data over the fence to the offline data warehouse world, said Bedos, but now online data is constantly being given to the offline data warehouse world, and the analytics teams are giving insight back.
In the process, Hotels.com is attempting to go from information, to predictive and prescriptive analytics. "It's not about telling customers what to do because we can't do that. It's about putting forward the right match to them, so we are making it so obvious to them that this is the hotel you should choose," said Bedos.
"In order to get there you need to have both the online and offline worlds talking together because a lot of the signals you get are not available in the offline world. It is about being able to understand what a customer is after," he added.