BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Luminar, PubMatic and Lotame Solutions Inc. are living proof that marketing and advertising are two of the killer applications for big data technologies like Hadoop, according to officials at the three firms, as well as IT industry analysts.
Advertising has always been about data.
director of platform strategy, PubMatic
Luminar, which delivers marketing insights about U.S. Latino consumers, decided to drop a traditional, relational data warehouse in favor of big data technology. PubMatic and Lotame, meanwhile, have teamed up to help publishers automatically evaluate and sell digital advertising inventory while analyzing data about the audiences those ads target.
All three companies will have to continue evolving, however, if they want to continue being successful in the future -- that is, if you follow the logic of Mike Gualtieri, a principal analyst at Cambridge, Mass.-based IT analyst firm Forrester Research Inc.
Firms that want to do a good job of reaching out to customers through digital advertising will have to analyze a great deal more than information about what those customers previously searched for or clicked on while surfing the Web, Gualtieri said in an email interview.
Going forward, advertisers will also want to look at -- for example -- Global Positioning System data to see where people have been, sales data to find out what they purchased and maybe even information about the movies they watched recently on Netflix. "Companies now wish to get all the data together in one place so that they can find deeper insights about their customers, in order to serve them like kings," Gualtieri said, "because if they are treated like kings, they will never leave."
Luminar drops traditional data warehouse in favor of Hadoop
Denver, Colo.-based Luminar, which specializes in helping consumer packaged goods companies, retailers and other businesses better understand the behavior and buying patterns of U.S. Latino consumers, turned to Palo Alto, Calif.-based Hortonworks for a Hadoop-based platform for crunching big data.
The company takes in information from more than 2,000 sources -- such as retailer loyalty programs, cable television invoices and utility bills -- and feeds it into Hortonworks' Hadoop distribution to gain a more distinct view of Hispanic communities than was previously possible through focus groups and surveys alone. For example, Luminar can tell its clients how Cubans in Miami spend money on technology, or how the much the typical Puerto Rican male in New York spends on food.
More on big data technology
Find out about data in the city
Luminar previously used a traditional relational data warehouse to run its analytic models, but decided last year to scrap that approach entirely in favor of Hadoop, according to Franklin Rios, the company's president. The company found that the traditional approach was too time-consuming and inflexible, explained Rios, who boasts many years of experience working with data warehouses. "My experience with data warehousing was a mixed bag. There were a lot of limitations," he said. "When I asked for something, it seemed like it would take a dog-year to produce it. And if you wanted more, the message was, 'It will cost you $5 million.'"
Moreover, warehouse development processes were slow and capital intensive, Rios said. "Once you adopted one, you were at the [vendor's] mercy because you were in proprietary environments." And moving from one data warehousing platform to another was like having "major open-heart surgery," he said.
Apache Hadoop's status as an open source standard and Hortonworks' ability to support the data management platform were important factors in Luminar's decision. Today, the company runs its own proprietary analytics models on top of Hortonworks.
PubMatic inks big data deal with Lotame
PubMatic is a Silicon Valley-based company that provides publishers with a digital media platform that automates the process of evaluating and selling advertising inventories. The company recently formed a partnership with New York-based Lotame, another firm that serves media companies with digital advertising space to sell. Lotame uses big data technologies like Hadoop, Hive and Pig to collect information and generate unique insights about publishers' audiences -- insights that ideally lead to better-performing advertisements and increased revenue.
Under the terms of the partnership, PubMatic's media management offering will offer full integration with Lotame's Crowd Control data management platform. The idea is to give users a single place to evaluate media inventory, as well as the audiences being targeted, according to PubMatic. The two companies say the partnership also will give publishers access to a wider range of information that can be used to simplify the process of audience segmentation and the process of creating customized advertising campaigns.
"Advertising has always been about data," said Richard Sobel, director of platform strategy at PubMatic. "[Even] when you go back to the '60s and the sponsoring of TV shows, it was always about finding an audience."