臺大管理論叢 NTU Management Review VOL.29 NO.1

Exploring Dual Business Model Choice of Brand and OEM Businesses 62 We used two primary data sources: archives and interviews. For each product, we began data collection by gathering archival data from internal financial and manufacturing departments. Internal financial department sources included 1) revenue generated from brand and OEM business, 2) dual BMs’ period, 3) the cost of goods sold with and without OEM business, 4) the operating income of brand and OEM BMs. From the manufacturing department, we collected first pass yield rates from 2008 to 2015. We began to use the data to develop chronological case histories for each product and thoroughly discussed their content elements, index selection, and how they can be compiled. This took around four months and resulted in documents over 10 pages long for each product, including key metrics such as market share, revenue, and profit. The second source of data was semi-structured interviews with internal and external informants located in Taiwan, Japan, the US, and other countries. Four to eight interviews were conducted per product line, with a total of 22 interviews from 2013 to 2015. To probe at suitable junctures, interviews were based on topic guides (see Appendix A). We interviewed the in-charge executive of every product line for 1.5 to 2 hours each. We also added conceptually consistent lay language to further explain possible capabilities and synergies. We matched data from interviews with archival material, to ensure triangulation. These first-hand interviews allowed us to create a historical reconstruction of incidents. Hence, we identified major related capabilities criteria and their associated synergies. We selected other internal informants based on these criteria: (1) long tenure of service in AV-Firm so that they can provide a temporal perspective on the product’s decision process, (2) direct involvement in the product line to provide first-hand knowledge, and (3) various functional/hierarchical positions, allowing us to obtain diverse views. Moreover, to complement internal informants, we interviewed outsiders including OEM buyers, channel resellers, and ex-colleagues via email, Skype, and face-to-face interviews. If conflicting interpretations of the same phenomena occurred, we clarified the discrepancies using emails and phone calls. By so doing, information could be triangulated and confirmed by several sources (Yin, 1994) instead of relying on informants’ memory, reducing individual bias and revealing complementary aspects of major decisions (Santos and Eisenhardt, 2009). Each interview lasted around 45 to 90 minutes and was tape recorded and transcribed. The first document contained unrestricted questions for informants to provide a broad view of each dual BM’s evolution and synergies. The second focused on