When international cloud computing/big data corporation EMC was looking to centralize and optimize data from across its global network, it turned to the team at Harte Hanks to build a trusted, accurate source of accounts and contacts.
Creating this optimized database of information from hundreds of international locations was part of a larger effort to develop what the EMC team calls “business analytics as a service.”
“What we are doing is enabling the enterprise with an analytical view of the customer for predictive modeling and analytics. Think of it as a large data lake that the enterprise has access to,” says Todd Forsythe, vice president, global marketing EMC.
The business analytics as a service platform pulls in data from all aspects of the customer experience and EMC’s internal ERP system, so it’s a very robust data set. Each business unit taps into this data lake, and builds its own analytic workspace to support their line of business.
Through the process of building out marketing’s view of this customer data environment, EMC integrated all of its marketing CRM data across all online and physical touch points. The EMC marketing team had a full view of customer experience from initial touch all the way through the sales cycle and closed bookings.
“As we looked at applying analytics to the business we felt that making dramatic improvements in data quality was critical to our ability to scaling analytics.,” Forsythe says. “Understanding an organization’s behavioral characteristics is critical, but then we have to link those characteristics to individuals so we can take marketing action. With that in mind, we contacted Harte Hanks to help build that data-quality bridge to dramatically improve the structure of the organization, and also the linkage of the contact and cleansing the contact record so it’s actionable and trustworthy.”
The end result is a system that allows EMS to view data from a global view, by country or by single customer—a master record that ties EMC sites together, as well as individual customers so they can be viewed individually or separately.
“EMC had a lot of systems, tools and people on staff to manage the data previously, but what was missing was tying all of those things together, creating standardized best practices and some methods and systems to really enable their global constituents,” says Brad Wamsley, managing director, Harte Hanks.
There were some challenges along the way when creating such a large optimized data set, the first of which had to do with ownership of customer data quality. EMC first had to ask itself who was really accountable for owning data quality on a site by site basis.
“Our view is that the individuals that are closest to the customer and closest to the sales experience should take accountability for ownership of data quality. Our plan was to put the systems and processes in place to maintain and sustain data quality in an always-on mode, then transfer the responsibility of maintaining data quality down to the country level,” Forsythe says.
“EMC field personnel in a given country now have the ability to look at our customer data, understand where there are gaps or challenges in data and take action on it,” Forsythe adds. “And now with the help of Harte Hanks there’s a high degree of confidence when they take action on cleansing or updating data in a real-time way.”
Another challenge was making sure the system would meet the needs of EMC’s global constituents by addressing things such as double byte character data for the Asian markets, and handling multiple languages in the hundreds of countries in which EMC does business, including Cyrillic language character records for eastern Europe and Russia.
When EMC created a team with its field organizations, which made sure all of the characters were accurate.
“We had the local language marketers and subject matter experts engaged in that process to make sure that what we were doing at a centralized level was accurate. The other task was creating a set of forensics to understand whether a source of poor data quality was based on a systematic error, a process error or an end-user error. What we found in our analysis was we not only had to address how we were managing hygiene around data, but that we had to make process improvements and focus on end-user education of importing data into the environment, just to maintain clean data,” Forsythe says.
When it came to making the system user-friendly for those interacting with it each day, the first goal was reestablishing confidence in the team that EMC has the capability to maintain an extremely high degree of data quality over time.
Over the years, our field marketers responsible for developing and executing marketing plans lost confidence in the global database. Our number one goal was rebuilding that confidence,” Forsythe says.
The second goal was streamlining the process so that it was easy for the end user to import data, but also be able to view the data so they could understand and have visibility into data quality.
The end result has enabled the EMC team to see analytics in action.
“We have built a marketing science lab, with big data analytic capability for marketing. Our ability to scale the output of any predictive modeling was based on data quality, so the data quality made an immediate improvement in our ability to scale analytics. Also, because we built a high degree of confidence that we have high quality data throughout the organization, the organization’s appetite to take action on the analytics was much stronger,” Forsythe says.
The ownership of customer data has clearly transitioned into the right hands and the field teams are taking ownership of data, identifying areas of data that are incomplete or inaccurate and cleaning that data up and reloading it into the system so it continually improves.
As a result, the action against the data is much stronger.
“There had been a lot of skepticism of any insight that was pulled out of the data because of data quality concerns, and a lot of that skepticism has eroded, which means we have truly become a data-driven organization and we’re embracing the insights that are derived by data,” Forsythe says.