Online gambling sector bets on business intelligence
Online gambling and gaming industries crave structured business data. Volumes can be huge and they need to react to customer behaviour
Although founded a decade and a half ago, with a business model belonging in the 21st century, WildTangent’s efforts at business intelligence (BI) have been beset by the same sort of problems older firms routinely encounter. The games distribution platform found that much of its business data was held in silos, with varying attention to quality, making it more difficult for business managers to reach decisions.
“We were in an environment where Excel ruled the day and the knowledge of the business was tribal," says Scott Moran, BI director at WildTangent. "It was handed down from one person to the other as they migrated through the company. That led to a fragmented view of how our business really worked.”
The outcome was a drag on business decision-making. “We wound up in business meetings with people having arguments saying ‘where does your data come from, I don’t trust your data’, before we ever got to the issue of the business,” Scott Moran says
Effectively gauging data
The firm, which also distributes games for social gambling, reaches 175 million consumers monthly across desktop, social and mobile platforms. A more coherent approach to business reporting was essential, Moran says.
Like any business, WildTangent wanted to know about revenue. And because the firm earns advertising revenue and licenses its advertising platform to other games distributors, advertising partners and platform partners need data to gauge how effectively they reach an audience.
The situation required a new approach to BI, Moran says. With a roll-out beginning in November 2011, WildTangent implemented a new technology stack, with a Tableau front-end application offering business users direct access to company data.
The firm relies on Hadoop clusters for Parsing website logs, alongside Hive databases help manage some of the workload.
It also uses Vertica databases (now part of HP) to integrate analytics data. The company uses instances of Microsoft SQL Server 2012 Enterprise Edition as the core database to run OLAP cubes and relies on Amazon Web Services (AWS) to enhance scalability via the cloud.
More on BI and online gambling
- Betfair takes online gambling to the iPad
- CIO interview: Martin Davies, head of technology at Bet365
- CIO video interview: Internet gambling operator imports IT skills for global expansion
Although use of Hadoop might suggest the project to be a big data initiative, the main reason for using the open source system was to manage large volumes of data, rather than unstructured data, Moran says.
WildTangent’s software ships in a large percentage of the PCs sold worldwide, so scale can be daunting.
“The amount of data that we have coming at us is astronomical," says Moran. "To be able to handle that, the only way to do it is with a distributed parsing system like Hadoop or Hive.”
As a result of the BI initiative, users now have more trust in business data.
“They can then focus on the business issues at hand,” he says.
Online gaming booms
Worldwide revenue from online games will reach $35bn by 2017, up from $19bn in 2011, a report from DFC Intelligence says.
Like other early entrants in the nascent online gaming and gambling industry, the founders of Gamesys Group – which operates online gaming sites including JackpotJoy and Sun Bingo – wanted to get products out as quickly as possible. The thought of up-to-the-minute reporting came later, says Nick Hughes, BI manager.
Initially, Gamesys took snapshots of data from live systems at midnight and loaded them into a management information database. The data would sometimes not be available to users until 5pm the following afternoon, by which time a working day had elapsed.
Meanwhile, business managers were reliant on the ad hoc SQL queries to extract meaning from the data. This left six analysts trying to satisfy many business users, creating bottlenecks, Hughes says.
During 2008, Gamesys began to roll out of strategic BI to get as close to real-time analytics as possible. Using a Cognos system supported by service provider Logicalis, the company populates a data warehouse with data replicated from its live systems using IBM’s Q Apply technology.
“We wanted to query live data without putting overheads on our live site as well,” Hughes says.
As well as up-to-date monitoring of business performance, the system helps Gamesys with compliance in this sensitive industry. “We are really proud of being on top of regulatory matters, but to do that we have to constantly monitor for and flag any suspicious activity, so we can close down accounts and stop taking transactions. If you only look at the data once a day, you could have lost a lot of money before you recognised a fraudulent account,” says Hughes.
The Cognos system offers business users access to Cognos Query Studio and key business KPIs using Cognos Metric Studio. This self-service BI model has freed up the BI team to assist the business in other areas, he says.
“Because of the alleviation of standard day-to-day queries, they get to spend time doing what they should be doing, which includes focusing on how to improve visibility rather than sitting there writing SQL queries all day,” he says.
Paddy Power’s need for agile BI
In the UK, online gambling reached £1.9bn in revenue during 2011, nearly double the £1bn recorded in 2007, according to Mintel.
Bookmaker Paddy Power comes from the traditional end of the market. It was formed in 1998 with the merger of 40 high street betting shops in Ireland. By August 2012, the firm had grown internationally and online revenue reached €191m – more than half of its €311m total revenue.
Yet, in terms of data management, it met with similar challenges other online gaming and gambling firms faced.
“We had a start-up culture. [In BI], we managed more than we thrived,” says Conor McMenamin, Paddy Power’s business intelligence manager.
However, the dynamism in the business made it difficult to justify the long-term commitment a data warehouse requires. “A lot of what we do is quite opportunistic. In those cases, by the time a data warehouse strategy delivers, the opportunities are gone,” he says.
To get around this problem, in 2009 Paddy Power opted to start using BI tools from QlikView on live business data in order to demonstrate their value to the firm.
“It gave the business confidence in what we could get out of BI and data warehousing,” says McMenamin. “Then we could work with the business to develop some quality product.”
Scalable platform
Paddy Power began the project in earnest during 2010. It was based on Microsoft SQL Server and SQL Server Integration Services with Excel-based clients. QlikView is one aspect of Paddy Power’s BI system that has survived from the old world to the new, McMenamin says.
“It was about putting in place a scalable platform that we could develop on quickly, and not be constrained by legacy issues,” he says.
As with other firms in the sector, the self-service BI model offered wider access to business data in a way that keeps BI professionals in control of data structure and governance, while giving them more time to focus on strategic goals.
The online gambling and gaming industries crave a structured approach to business data. Volumes can be huge and business need to react quickly to changing customer behaviour. The drive to set the right odds, protect against fraud and monitor audience behaviour all require a sophisticated technology and management. It's odds on that BI skills will be in high demand for some time to come.
