Hays handles rapid data growth with Google Search Appliance and Oracle BI tools

Hays has chosen Google Search Appliance and Oracle business intelligence tools to help gain more value from its growing mass of data.

Share

Hays has chosen Google Search Appliance and Oracle business intelligence tools to help gain more value from its growing mass of data.

International recruitment firm Hays has around 7,800 staff across 33 countries, including 2,000 consultants, and is continually looking to expand its business internationally.   

At the heart of the firm’s business is the data provided in candidate CVs, which can be searched to allow consultants to quickly match the relevant candidate with a job, in a highly competitive recruitment market.

In total, Hays stores 10 terabytes of information in Oracle 10G databases, with approximately 35 million CVs on its records.  The company hosts its data in a Manchester and Leeds data centre run by Computacenter, dedicated to servicing the UK market, as well as other parts of Europe.  A separate Asian data centre is also used to service other regions.

According to Hays CIO Steve Weston, the number of CVs, and hence the amount of data, is rising rapidly each month. 

“We receive around 300,000 CVs a month, around 4 million a year,” Weston told ComputerworldUK. “That is a lot of data being received all the time - it doesn’t take a genius to work out that is growing really quickly.  All of that data needs searching, needs insight and needs work doing with it.”

One of the problems faced by Hays in the past has been the inability to access information in siloed databases in different regions, with up to 100 separate databases in each country for different industries.

As part of an IT transformation project, Hays decided to implement a new tool to allow federated searches across its numerous data bases, and enable staff to search unstructured data using a common platform for all geographies.

Hays had previously used the search functionality it had developed within its modified version of Bond Adapt 8 recruitment software system.  However the basic system created a number of problems, such as slowing down consultants who needed to know specific codes to perform searches, and created a large workload around categorising and manual indexing of the vast amounts of CVs accrued by the company.

This led to a decision to implement Google Search Appliance 7.0 during an IT transformation project which lasted up until 2010, installing one appliance for the UK, one for Europe, and one for Asia.

Weston explained that the decision to implement GSA was attributed to Google’s reputation in search, and its commitment to continually developing its search solutions. This meant turning down similar solutions from Oracle and Autonomy.

“We looked at all the major players.  Our decision was to go with the company that offered the best integration, and fastest search that would be pertinent to the recruitment industry.”

“We went through a normal evaluation process.  The GSA was much more relevant on a number of dimensions; it was relevant because there was no real training we had to do as the tools are familiar, it ‘plugged and played’ into our architecture very simply, and it was relevant in the way that we could scale it and federate it around the world.” 

The system now integrates with the latest, heavily configured version of Bond Adapt, the core database system running on top of Oracle 10G, and an automatic CV parsing software system from Daxtra.  The tool also integrates with a contracts database and a compliance checking system, as well as Broadbean’s job-board posting system.

Weston said that Hays consultants can now search across all databases in different industries and across a number of regions with federated search functionality.  This makes it easier to match candidates for roles in global industries such as oil and gas, or financial services which can involve recruiting candidates outside of the reach of the UK databases.

GSA also now allows Hays staff to perform much more in-depth searches across all of the company’s databases, using unstructured data.

“We also built in more unstructured data. Historically the recruitment industry has used very structured, codified data.  Now we do a lot synonym type searches, as well as Boolean searching with the development of a new front end using GSA,” said Weston.

The system has since been updated with features including Google’s automatic translate system, to make it easier for consultants to view international CVs. 

The processing power of each of the three bright yellow appliances boxes has also been increased to deal with the continued growth in data.

In addition to its expanded search capabilities, Weston also noted that around 18 months ago Hays introduced business intelligence tools to help gain greater insights into its masses of data.

“The other challenge that the recruitment industry has is around the big data issue,” he explained.

“We have built comprehensive warehouse and business intelligence stuff around this.  We are able to drill down at all levels of the organisation, and see, say, ‘how many accountants have we seen in the last few days? Where did they come from, where were they educated, which job boards did they apply from?’  All of that type of data we can now analyse really quickly.”

This involved implementing the Oracle Business Intelligence Enterprise Edition (OBIEE) system, he said.

“We use that as our data warehouse interrogation, big data toolkit.  That is all about turning data into insight.”

“Everybody uses it, from the consultant doing his job, right up to the chief executive.  It is the same tool used by everybody to get the big data questions answered. The whole system has allowed us to have a huge difference in the way that we do things.”

Find your next job with computerworld UK jobs