Beth Sexton, senior project manager, and Sathish Gaddipati, director of Enterprise Data Management and Business Intelligence (BI) at InterContinental Hotels Group


Case Study

Room to grow

IHG boosts information delivery performance with a data warehouse appliance.

When IHG (InterContinental Hotels Group)—operator of more than 4,400 hotels worldwide—outgrew its data management system, company leaders turned to Teradata to solve myriad data-processing speed, volume and performance issues. The result positioned the company to meet aggressive near- and long-term growth, profit and improvement goals.

Executive summary

The company: IHG (InterContinental Hotels Group) operates more than 4,400 hotels with 620,000 rooms in more than 100 countries. With seven brands—including Crowne Plaza, Holiday Inn and Candlewood Suites—the company welcomes 160 million guests each year. Additionally, IHG boasts the world’s largest hotel loyalty program, Priority Club Rewards.

The challenge: In a three-year span, the number of enterprise data warehouse (EDW) users doubled, the number of data sets tripled, and the company was getting inconsistent query results. The legacy EDW had reached its scalability limits.

The solution: IHG implemented a highly scalable data warehouse appliance from Teradata that allowed IHG to easily forklift existing EDW data to the new platform, which improved query performance speed and accuracy, and integrated easily with existing business intelligence (BI) tools.

The results: Queries that used to take IHG users hours now take minutes. Additionally, the organization can create a 360-degree view of its customers to improve marketing campaigns and boost loyalty programs.

Amid the final stages of its initial Teradata Data Warehouse Appliance implementation, IHG leaders are already reaping the rewards. Teradata Magazine recently spoke with IHG’s Sathish Gaddipati, director of Enterprise Data Management and Business Intelligence (BI), and Beth Sexton, senior project manager, about how their organization came to select Teradata for its data management needs, their implementation journey and the results to date.

Q: What challenges or pain points led you to upgrade your data management system?

Sathish Gaddipati: We migrated from an entry-level data mart to an enterprise data warehouse [EDW] in 2004. Since then our data and number of users have doubled, while our data sets have tripled. The legacy system simply couldn’t accommodate the growth.

Additionally, our tables used special indexes. When we loaded data at night, we dropped the special indexes and there would be an impact on query performance for the users of the system in the late night or early morning. Ours is a 24x7 global data warehouse where we have people in the Philippines, the UK and India who were not very happy with that process.

The other issue that we had was the inconsistency of the queries. The whole environment was not really conducive to analysis. It was time for us to upgrade the infrastructure.

Q: Why did you ultimately choose the Teradata Data Warehouse Appliance?

Beth Sexton: There’s a risk in switching vendors and laying out a platform for a huge new project with visibility across the company. There is also a risk in changing platforms, which certainly had to be considered. However, we were extremely impressed by the speed and performance of the Teradata Data Warehouse Appliance.

SG:The major decision factors were scalability, query performance, easy integration with other EDW and BI tools (Informatica, SAP BusinessObjects, SAS, et cetera), data integrity, our ability to forklift the current EDW, and vendor track record. But I think total cost of ownership was the main driver.

Figure: IHG architecture

Click to enlarge

To determine total cost of ownership, we evaluated migration, product cost, and regular and annual maintenance, including how many database administrators [DBAs] were required to maintain the environment. Teradata was number one of all vendors. It’s the market leader.

Q: How long did your search for a solution and implementation take?

SG:We identified the need to upgrade our solution in 2007. We began looking for the options in the market in early 2008 and made the decision to go ahead with Teradata in September 2008. The core data was migrated in four months, from December 2008 through March 2009. All of the systems ran in parallel for almost four months to make sure everything was in place before we shut down the existing EDW on July 15. Our whole schedule was driven by a strict statement of work, with a project plan and defined milestones, that enabled us to complete the implementation on time and within budget.

Q: Was the entire process harder or easier than you expected?

SG:One reason we selected Teradata is for the experience its professional services team has in facilitating a migration. During the forklift, there are three types of work. First is the database work. Second is data integration. Third is the reporting work, where we needed to map our reports to the new structure and make sure the data was good.

For the first and third tasks, it was much easier than expected. We thought they would be very complex, but they weren’t. The middle component was exactly the way we expected.

I think our data migration strategy is the thing that saved us. We made incremental copies of our data so we would always have a reasonably current mirror image of our work. That way, if there was a mistake, we could go back and fix it quickly.

BS: So in a worst-case scenario, you only had to apply seven days of data, rather than 30 or 90 days.

"We could give a complete analysis within a day to our senior management, who appreciated that we could respond to their business query so quickly with answers and with various possibilities."

Q: What benefits have you realized from your new system?

SG: Query performance has increased dramatically, from hours to minutes. The change was so dramatic, in fact, that senior management didn’t believe our results at first. Once we were showing a consistent response time for a couple of weeks, then everybody started believing that this is not a one-time thing; we can produce substantial results. Our senior management was quite pleased with the results. They were very happy.

Also, one of our biggest problems with the previous system was with customer segmentation. We segment the customers based on the reservation data and stay data. For example, you take a guest and look at how many times he’s stayed, whether he’s choosing the same brand and hotel, whether he travels to a particular destination or goes to different destinations. We were never able to do that before, but now we are doing customer segmentation every day.

BS: On a corporate level, our business partners see enterprise data and business intelligence as a business enabler, and we are well positioned for new corporate initiatives. Additionally, we have an increased capacity for campaigns and promotions. Also, it’s very easy to maintain the new platform. Previously, we needed to do a lot of query tuning and had three resources working on it. Now we have one DBA who takes care of the query tuning as well as administration, which results in lower operating costs.

Behind the solution:
InterContinental Hotels Group

Database: Teradata 12

Platform: 2 Teradata Data Warehouse Appliances:

  • 8-node production system
  • 4-node disaster recovery system

Teradata Data Mart Appliance quality assurance system

Users: 2,350 (100 concurrent)

DBAs: 1 plus

Data model: Logical—Erwin
Physical—Star Schema/Dimensional model

Operating system: SUSE Linux

Storage: 40TB total for all 3 systems

Teradata utilities: Teradata Tools and Utilities: FastExport, FastLoad, MultiLoad, Teradata TPump, BTEQ

Q: Can you tell us how your new system helped IHG make critical real-time decisions at the beginning of the H1N1, or “swine flu,” outbreak?

SG: Well, we wanted to know how swine flu had impacted our bottom line and how it might impact it in the future. In the past, it used to take a few days to run similar queries. Now we can project something like this in minutes or hours. So, we could give a complete analysis within a day to our senior management, who appreciated that we could respond to their business query so quickly with answers and with various possibilities.

Q: What role does Teradata Customer Services play in your new implementation?

SG: We have three different service plans in place, so if we have any problems, the appliance dials and sends a ticket to Teradata Customer Services, and they fix the problem. So from our perspective, we manage only the data objects. The complexity within the environment is handled by Teradata. That is the biggest advantage over what we had.

BS: That’s right. The Teradata team provides a single source for phone calls and repair work. They deliver a total solution.

Q: What will the future hold for the system?

SG: We have plans to bring in our finance data so we can do side-by-side analysis between operations, marketing, sales and finance data. We also plan to bring in our rates data, rewrite our loyalty program and build our master data management discipline for the sales organization.

We’ve got the foundation; now we are going to work on top of that. We are very optimistic of meeting those goals. Within one year, we hope to achieve all of our strategic objectives.

Q: How was the Teradata team to work with during this ongoing journey?

BS: There’s been some give and take. At the end of the day, we always come back together, and we’re moving forward. It’s been a good relationship, and we don’t expect anything less in the future. We’re excited about the platform. Our business users are excited about the platform, and we feel equipped for these new initiatives and for the future.

Your Comment:
Your Rating:

Fuzzy Logix