Baku, May 1 (AZERTAC). Emerging data-mining software tells whether employees are helpful, toxic, or management material—and how well the company functions.
Consider your workplace e-mail style. When asked a question, do you confer with others and attribute your response to the group? Do you avoid making a decision in case you might need to reverse course? If so, you may be a buck-passer, causing a productivity drag in an organization.
In the latest in digital employee monitoring, new software can identify such people and discern a range of other traits and behaviors, potentially allowing management to intervene or assign people to tasks better suited to them, boosting productivity.
Software from Cataphora of Menlo Park, California, sorts through terabytes of data created at companies and government organizations—e-mail, instant messages, calendar events, documents, and even phone logs—to pull together a digital character profile of not only individuals but the organization as a whole.
For example, Cataphora tracks use of exclamation points, font color, capitalization, punctuation "cursing," the way people sign off in an e-mail, and the overuse of certain words, such as "please."
The tool—and associated visualizations of the data—builds on the company's expertise in providing e-discovery and analysis services for law firms poring over company data; Cataphora is now commercializing the monitoring technology.
The idea is to expand the concept of digital monitoring beyond traditional efforts, such as checking whether employees are visiting porn sites or making personal phone calls, to produce a deeper understanding of employee traits and organizational health, says Daryl Nord, professor of management information systems at Oklahoma State University. "No longer does discovery and monitoring software simply view, store, and report," says Nord. Rather, it is able to "detect anomalies, identify changes over time, or locate different work habits among peers who perform the same job functions."
© Content from this site must be hyperlinked when used.