This study examines the duration of brand alliances (or co-brands), a brand strategy in which two brands are offered as one joint product. Previous research has suggested these alliances are short-lived, but little empirical evidence exists to explore what may drive the longevity of such alliances. The study uses actual market data for 524 brand alliances in 83 product categories of consumer packaged goods during a 13-year period. Controlling for market share, several factors that might influence brand alliance duration are examined: type of alliance, ownership by the same parent company, and the number of relationships. Using a Bayesian hazard model to estimate duration, the results show a brand alliance lasts longer if it is an ingredient brand alliance (rather than licensed) and both brands are owned by the same parent company. Additionally, having more partnerships helps a brand alliance last longer, but too many alliances have a negative effect on staying in the marketplace. The findings suggest that brand managers looking to enter into a brand alliance can anticipate how long the product might last based on these partnership factors.

Introduction

A growing strategy among managers is the use of brand alliances, or co-brands, which feature two brands in a single product (Keller 2013). Chiefly, the primary brand imports the brand equity of a secondary brand to offer a unique value proposition to consumers (Uggla 2004; Koschmann and Bowman 2016). Brands are a key asset of firms as a resource for obtaining competitive advantage (e.g., Grant 1991); strategic alliances represent the collaboration between two or more firms to share resources or information (Besanko et al. 2007) and have become ‘‘common and important structural vehicles for business development’’ (Albers et al. 2016, p. 582). Indeed, the practice of brand alliances has grown from 3.5 to 6% of all new consumer packaged goods launched in the USA (Schultz 2014). Although brand alliances are of growing importance to managers, less understood is how long these alliances last in the marketplace. Whereas prior research has taken an internal approach of the firm regarding alliance duration (e.g., Harrigan 1988; Park and Russo 1996), few studies have approached the issue from a market performance aspect. Indeed, only scant research has explored market effects of brand alliances (e.g., Desai et al. 2014; Koschmann and Bowman 2016; Swaminathan et al. 2012). More recently, research has found that brand alliances reduce the brand equity of the primary brand (Koschmann 2017). While these studies have emphasized the outcome, another aspect of brand alliances is the process and duration. This study makes three contributions to the brand alliance literature. One, the research proposes that the type of brand alliance, parent company ownership, and the number of brand alliances have an effect on how long a brand alliance persists in the market. Drawing on brand strategy research, it is proposed here that ingredient brand alliances (where the fundamental product composition is changed: Keller 2013) last longer than licensed brand alliances. Additionally, when both the primary and secondary brands are owned by the same parent company the brand alliance survives longer. Furthermore, if the primary brand engages in more brand alliances, this should lead to longer alliance durations; however, too many alliances will hasten any single alliance to exit from the market. Second, whereas prior brand alliance research has often used consumer survey data (e.g., Decker and Baade 2016; Dickinson and Heath 2006; Simonin and Ruth 1998; Van der Lans et al. 2014), the theories are tested using large-scale aggregate field data of consumer purchases. The data cover the market performance of 524 brand alliances across 83 product categories of consumer packaged goods sold in the USA during a 13-year period. Third, the data are examined using a Bayesian hazard model. The findings show that most brand alliances are short-lived; the median duration is 2 years. However, this average masks some issues of interest to managers. Brand alliances that are licensed in nature (vs. those that utilize the secondary brand as an ingredient or component) last much shorter, on average. When the parent company owns both brands, there is no significant difference in survival when compared to brand alliances that are not owned by the same parent company. Finally, the number of brand alliances a brand engages in is more likely to keep any particular brand alliance, on average, in the marketplace. However, too many alliances have a negative effect on keeping any particular brand alliance in the marketplace. This article is organized as follows: First, theory is developed around the types of brand alliances and their durations. Then, the theories are tested using a Bayesian duration (hazard) model for continuing in the marketplace. The research concludes with a discussion of the empirical findings, implications for managers, and avenues for future research.