IBM Working On Analytics Suite For Advertisers
IBM has been developing code that it will use to build a suite of analytics products for marketers and advertisers. The suite of applications will consist of modules -- off-the-shelf software for IBM clients as part of a service package that chief marketing officers (CMOs) can customize to fit specific needs. Big Blue will offer services for the analytics package supporting campaigns, Michael Haydock, IBM global lead for business analytics and optimization, told MediaPost.
Haydock, an operations research scientist for IBM's customer analytics practice, has been writing a strain of linear programs for asset allocation, allocating resources into customer groups and advertising media using multi-objective optimization, where customers have multiple objectives. He also wrote the original multi-dimensional clustering codes.
Using statistical methods and measures, IBM continues to build a practice supporting advertisers and marketers. The company has a small group of people called a "practice" to support statistical analysis for large databases filled with 50 to 100 million clients.
Supported by this code, Haydock leads a worldwide team of experts helping to "generalize" code. Efforts will help build an asset-based platform through logic, statistical analysis and reporting tools the company has acquired throughout the years. The suite has not been scheduled for release, but the company has been working long hours to develop and launch the project soon.
The suite of five modules will collect data, cluster information, forecast demand, and suggest the best media to place advertisements, tools that optimize and ingest content, and reports that analyze content based on performance.
Haydock is not concerned about IBM going up against companies like Omniture and Webtrends. He says IBM has the technology and skill to crawl the Web and analyze massive amounts of data to find the correct media mix. It also has a massive research division.
The suite will support people who advertise a lot and want to make smart investments. "No matter how much the company's budget for advertising, it's never enough," Haydock says. "It's all about getting the correct mix in television, online, radio and print."
The team has been working with CMOs to allocate advertising and marketing resources to help get media buy allocations correct. One IBM client relies on 19 advertising methods. The goal -- to predict customer response for each by creating clusters to determine those who respond better to one media versus another -- has been in the works for 18 months.
Acquisitions that support the transition include Coremetrics in 2010, and SPSS in 2009.
The modules will help advertisers answer questions -- for example, does television drive search? "It's a puzzle that needs to be unwound," Haydock says.