Authors: Bradley R. Staats (University of North Carolina at Chapel Hill), Melissa A. Valentine (Harvard Business School), and Amy C. Edmondson (Harvard Business School)
Publisher: Harvard Business School, Working Paper No. 11-031
Date Published: December 2010
As more and more companies use virtual or ever-changing teams to tackle projects, it has become crucial for employees to avoid reinventing the wheel every time a new problem arises. Accordingly, gathering and storing an organization’s knowledge for easy employee access has become big business: In 2007, U.S. companies spent about US$73 billion on knowledge system initiatives. But does being able to access that codified knowledge — from white papers, case studies, or documents that provide step-by-step procedures — actually help teams perform better? According to this study, teams using well-stocked databases may become more efficient, but they often don’t produce higher-quality results. There are ways to structure teams’ access to databases, however, that help them get the most out of their searches.
The researchers combed internal records from Wipro Technologies, a global software firm, to see how project teams used the company’s stored knowledge. The records contained objective performance outcomes on about 300 of Wipro’s software development projects as well as information on more than 9,000 members of the project teams. Use of the knowledge database was tracked on a per-click basis, allowing the researchers to study the relationship between access and performance on a “fine-grained level.”
The study first focused on overall usage of the knowledge base and its effect on performance, examining such factors as frequency of use, team size and experience level, and types of projects and their complexity. In general, productivity, in terms of saving time and money, was higher for teams that had access to the knowledge base. But the quality of the work (as measured by the number of defects in the software code that the teams produced) typically did not go up in a significant way for teams that used the database. The reason, the study says, is that improvements in many projects require an integration of knowledge that goes beyond the general database search for discrete information.
However, when teams were geographically dispersed, the opposite was true. They became less efficient, because the team members were largely on their own in learning how to use the database. At the same time, dispersed teams that relied on the database made fewer coding mistakes than similar teams that had no access. In this situation, the database became the glue for the team members, helping them to integrate their knowledge and resulting in a boost in quality.
Only when teams face constantly changing tasks does repository use result in both better efficiency and better quality. When the project changes suddenly, teams may require knowledge that only a repository can provide quickly and comprehensively. For managers of teams whose projects continually evolve or are subject to sudden change, ensuring access to a database can be crucial.
The researchers also explored how performance was affected by the two basic ways that groups structured their access to the database: either concentrating the search effort in the hands (and heads) of a few members or distributing it more evenly across the team.
Using a “divide and conquer” strategy (in which a few team members become experts at navigating the database and learning its quirks) will make the team work faster, the researchers found, but could lead to reduced quality — for example, one person interpreting the data could make a mistake that could harm the entire project.
Alternatively, a “share and share alike” strategy (in which all team members have access to the database) could boost the quality of the project because employees who do their own searching and downloading understand the data better and could therefore propose their own novel solutions. But that approach takes much more time.