The MapR big data platform enables Harte Hanks to integrate multiple new diverse data sources to enrich and enhance customer databases. The result is Harte Hanks can offer its clients faster and more accurate customer insights and more complete customer profiles so they can create smarter, more relevant and effective customer interactions.
Harte Hanks evaluated the three leading Hadoop distributions and selected the MapR platform for their needs. “Multi-tenancy, data mirroring and the UI were key points, but the tipping point was multi-tenancy."
Harte Hanks, based in San Antonio, Texas, is a marketing services company focused on developing better customer relationships and experiences and defining interaction-led marketing. Their approach with clients starts with discovery and learning; which leads to customer journey mapping, creative and content development, analytics and data management; and ends with execution and support in a variety of digital and traditional channels. Clients include Sony, BMW, Adobe, Kohl’s, Panasonic, US Airways and Volvo.
Harte Hanks wanted to increase the quality and speed of the customer-targeting services they offer their clients. To do this, they needed to be able to enhance and enrich their client’s contact databases with more social and digital data sources. They knew that trying to integrate new data sources into their traditional Oracle database would be slow, expensive and could not scale with the business as it grew. So they turned to big data and Apache Hadoop.
Harte Hanks evaluated Cloudera, Hortonworks and MapR distributions of Hadoop and ultimately decided that MapR was the best platform for their needs. “Multitenancy, data mirroring and the UI were key points, but the tipping point was multi-tenancy. That helped us make our decision,” says Sean Iannuzzi, Head of Technology and Development, Harte Hanks.
The MapR data platform enables Harte Hanks to enhance the performance, scalability and flexibility of their solutions so clients can more easily and quickly integrate, analyze and store massive quantities of data without impacting performance. “This new solution enables our clients to integrate all kinds of digital data, survey data, reference points and more, all while maintaining the performance and ease-of-use they’ve come to expect with us,” says Iannuzzi. Harte Hanks also uses Apache Drill to quickly analyze structured and semi-structured data to shorten the time to value from data.
Harte Hanks is seeing multiple benefits from MapR including higher quality data, deeper consumer insights, faster turnaround time to clients, operational cost savings, and a scalable platform for future growth.
More data yields higher accuracy and new customer insights. While integrating more data sources into client databases provides higher quality data, it is very challenging to do with traditional database technology. “We wanted to load large sets of data as quickly as possible while maintaining performance capabilities. With traditional technology, you have to keep increasing the capacity of your servers to do this, and that’s too expensive,” he says.
Harte Hanks used to be limited in the amount of data they could import. For example, if the data source had 500 fields, they would only take a subset of the most appropriate fields. “The customer would usually come back and ask for other fields and we’d have to go back and load those. There was a lot of inefficiency and it took a long time,” he says.
“Now we just load all the data. We don’t throw anything away. This is a totally different approach to how data is loaded and processed. We save time because we don’t have to keep reloading it. We just change the queries,” he explains. “And the data quality goes up significantly.”
Multi-tenancy capabilities cut operational costs. The unique MapR multi-tenanc y capabilities provide significant advantages for Harte Hanks. “Other distributions of Hadoop would be more expensive because they would require individual clusters per client,” he says. “MapR allowed us to leverage the same investment in infrastructure for multiple clients without having to create multiple clusters.”
“With a traditional database solution, we would have to purchase larger, high-end machines at $50-100K a piece to scale as we expanded our business. Now we can buy commodity servers for a few thousand dollars and get the same performance,” he says. “There is a big cost savings over time.”
Performance accelerates turnaround time to clients. Harte Hanks relies on the MapR high-performance solution to respond to their clients faster. “Our business depends on how quickly we can process data,” says Iannuzzi. “Processing a report based on 500 million records with our old system would typically take one to three days. With MapR parallel processing capabilities, we can do the same task in minutes or hours.”
This is especially important when a client is processing data from a holiday or shopping event like Black Friday when they need to act on the data rapidly. “Our clients are able to act more quickly on the data and make better decisions to market to their customers,” he says.
Integration and user interface ease development and management. Harte Hanks also appreciates the overall MapR user interface and integration with ecosystem components such as Apache Drill. “It’s well integrated. MapR provides a layer on top of Hadoop that makes the way data is stored transparent. The visual and logical grouping is very easy to set up and made the development process very easy,” he says. “From a maintenance and management perspective, having ever y thing in one location makes centralized administration much easier.”
Scalable, flexible platform for future growth. MapR provides a dynamic and configurable solution that flexibly aligns with other Harte Hanks customer initiatives. The framework allows them to handle the growth of data over time, enabling them to scale out vs. scale up. “The biggest benefit is we’re able to process all of the data we receive in a timely manner. If we need to increase performance or scale, it’s almost plug and play because of the architecture and how the software and infrastructure deployed,” he says.
MapR supported platform provides stability. Harte Hanks knows that having a big data solution gives them a competitive advantage in the marketplace. “With the growth of data over time, we would be behind the times if we didn’t shift to a big data solution,” he explains. “But without support for that platform, the big data solution wouldn’t be successful. Options are limited with other deployments.”
“MapR has extended Hadoop to provide more benefits, and stands behind the technology,” he continues. “If something were to go wrong, we can rely on MapR to address it. This provides a level of stability with the platform that we need."