Three mark factors to realize large data mining

Many companies are exploring how to tap into the value of big data, but one type of company has to be good at doing it: data analysis providers. The business model of these companies determines that they must be able to effectively and consistently provide the data products and services that customers are willing to pay for.

Here's a great data analysis provider-ComScore, how to tap into the value of large data through 3 key factors (a convenient, scalable platform, an employee team that understands data analysis, and a deep understanding of customers).

ComScore provides services based on 14 PB of multiple sources of online data, which are collected in real time from a global scale, what companies now call Big data.

The source of the data has 4 main channels. The first source is sample database data from 2 million Internet users, with 1 million people in and outside the United States. The sample library member allows comscore to secretly collect user behavior and demographic characteristics. The census data is the second source, from sensors that have been allowed to be placed in about 90% of the US digital media agency's hundred companies. A third source of data is the use of proprietary surveys to obtain cognitive data from members of the sample library. The fourth source is data obtained from strategic partners. For example, ComScore uses membership card store data to help customers combine online advertising activities with offline shop buying behavior.

In the late 90, the mainstream business technology solution that managed the ComScore data factory had not yet emerged. ComScore has developed an efficient, patented technology platform to store and manage large data.

By 2013, the technology team had transformed the platform from a proprietary solution into a service-oriented architecture that supported 3 critical systems running on MAPR Hadoop and pivotal Greenplum databases, each with separate tasks and extended requirements. ComScore needs to stay on top of the platform uninterrupted, and over the past 12 months the volume of data has increased by about 80%.

ComScore relies on its superior staff to master the art and science of large data. It has grown into a global team of 1200 people, each with a different level of "data scientist" capabilities. In order to keep pace with the growth of the company to choose and develop these employees, the company led the implementation of a number of talent management strategy.

One strategy is to recruit and analyze talent from the university's business schools and math majors, and to provide more analytical training to her. ComScore has an ongoing in-house scope training program at-comscore University, where most of its staff are involved.

Another practice is to develop a broad, deep understanding of the data analysis of the work team, according to the different skills of employees to set up different groups throughout the company, but at the same time, employees must have sufficient to ensure that these teams can communicate the basic skills. The teams that need to communicate often work in the same place.

The leadership of the company also uses a matrix organizational structure to combine talented people with different skill sets. For example, comscore runs several parallel teams, each with a business product manager to understand the roadmap for product development, and analysts to monitor data quality and analytical methodologies. This approach is considered to be particularly important for product development processes.

ComScore publishes about 150,000 online reports a day. The challenge is how to spread good reporting so that customers not only gain insights but also combine this insight with work tasks. When customers put their insights into action, such as changing the strategy of an ad campaign or positioning a new customer segment, they can derive value from ComScore's products and services. But as a provider of data analysis, ComScore lacks control over how customers use data.

What comscore can control is its own understanding of customer needs. The company is committed to digging deep in the field of expertise and understanding the challenges its customers want to address. With this knowledge, ComScore makes big data "consumable" and can proactively help customers identify which insights are actionable. " Admittedly, data can be overwhelming in a short time, motivating customers to focus on solving key issues and then applying them over and over again.

ComScore provides its customers with software tools with graphical interfaces, consolidated visualization and graph rendering, supports customer self-service through extensive use of wizards and template reports, and provides real-time access to data through the control Panel when immediate decisions are needed, as well as creating knowledge portals to support knowledge sharing. Further, the company has set up an organizational unit to provide support for the consumer of its products and services.

(pictured, ComScore's typical client)

In the space of large data, we must constantly adapt to keep pace with its rapid development.

Many years ago, ComScore had only expected to measure the number of visitors to a Web page with a client's ads, and today it must report whether the ad was actually seen. In the early days, corporate clients had little interest in non-US data, and the global development spawned market demand for ComScore to cover global data.

In addition, ComScore needs to follow up on technology-related changes. For example, customers may need to analyze bursts of data types such as data streams from video and smartphone. Another important example of change is the continued growth in the number of users accessing online content using a large number of different devices. In this way, corporate customers need to be able to identify independent users across multiple devices.

ComScore's platform, talent and deep customer understanding gives the company enough energy to cope with the vicissitudes of the big data-industry environment. Its convenient and scalable data platform enables comscore to adapt to the rapid growth of data volumes and to obtain new data types. The combination of teams and skills, combined with a team of working teams with deep analytical skills, enables it to adapt smoothly to changing requirements. Good training for customers using analytical tools allows the company to anticipate future trends and adjust its products and services accordingly. CISR (MIT Research Center for Information Systems) believes that the ComScore approach can be considered for good service with companies interested in using large data.