Over the past few decades international publishing firms and larger research libraries have accumulated countless digital objects, housed throughout the organization and rarely with metadata to facilitate retrieval. [A digital object here is any identifiable, organization-specific file (an image, chapter from a book, part of a published article, shoulder notes, chapter headings, etc.).] But as costs, competition, and consumer needs change (Iansiti et al., 1999; Clegg et al., 1997; Moad, 1994; Suwardy, Ratnatunga, and Sohal, 2003), publishers look to “repurpose” digital objects by altering work flow or modifying enterprise-wide information systems. To gain the advantages of repurposing, however, is no easy feat. In a wide survey and case study report, Clegg et al. (1997) found 80-90% of IT investments do not meet their performance objectives, and that less than a quarter of new technology projects achieved the integration of technology and business goals. Furthermore, 80-90% of the issues companies encountered when implementing new IT are organizational, rather than technical. As other researchers made clear, technology does not exist by itself within an organization, and many other factors contribute to the success or failure of the new technology implementation (Sambamurthey and Zmud 1999).
Seven case studies conducted in 2008 at four international publishing houses, two large university research libraries, and one museum suggest similarities of need and similar difficulties. All cases use some kind of commercial content management system and have sought similar in-house solutions only to arrive at several roadblocks. This study identifies the roadblocks at the management, organizational knowledge, organizational communication, and technical levels.
In these cases, repurposing offers the organization a chance to redeploy its assets and rethink its strategies. The concept of “user” is expanded to include interoffice communication as well as external communication to clients. Top-down efforts in each case failed to achieve the goal because of a lack of understanding of metadata practices, fear of loss of control over objects, commitment to large capital investments even when shown not to be able to achieve the goal, and considerable resistance to changing existing work behaviors, which echos McFarlan's (1984) and Proctor and Doukakis (2003) reports. In-house retraining in metadata standards and especially issues associated with interoperability, data conversions, and probabilistic classification techniques theoretically will alleviate many concerns, but management techniques (top-down management imposing poorly-understood recommendations from outside consultants imposed at the vice president level) contributed to resistance (cf. Erdogan et al. 2008; McFarlan, 1984; Erodgan 2008; Ford, Ford, & McNamara 2002; Jarret 2004).
A partial analysis of the case studies brought several recommendations. Attempts to integrate the candidate solutions magnified the key problems and brought to the fore in each organization weaknesses, especially at the executive, middle-management and technical services levels. Ultimately in all cases, a proof-of-concept technical solution was required that demonstrated support for capital investment, staff emotions, interoperability across heterogeneous data models, and current work flow, while dramatically altering and updating storage and retrieval of digital objects and human-computer interactions to integrate the objects into staff work flow, among other issues. Of particular interest at the executive level was the potential of probabilistic automatic classification of digital objects and the need for finer-grain descriptive and administrative control over objects. Middle managers' roles become important in the implementation process as they not only help staff with sensemaking as changes take place (Balogun and Johnson, 2004), they are also are more likely to be able to provide emotional support while the impact of the technology was unclear (Huy, 2002).
The project concludes with a series of recommendations at the executive, middle-management, staff, and technical staff levels.