A Unified Online-Offline Framework for Co-Branding Campaign Recommendations
Xiangxiang Dai, Xiaowei Sun, Jinhang Zuo, Xutong Liu, John C.S. Lui · 2025
Read A Unified Online-Offline Framework for Co-Branding Campaign Recommendations by Xiangxiang Dai, Xiaowei Sun, Jinhang Zuo, Xutong Liu, John C.S. Lui (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-22254.txt.
Introduction
Xiangxiang Dai and co-authors gave co-branding its first systematic treatment in 2025, as a recommendation problem with money on both sides. Two brands teaming up can widen each other's reach, but picking the partner is hard: budgets are uneven, willingness is uncertain, and markets move.
Their framework joins an offline graph with online learning. The graph links initiating and target brands and scores the promise of each pairing. Live feedback then updates the scores while the system balances trying new pairings for long-term gain against milking proven ones. High opening costs for new partnerships get their own mitigation inside the design. Readers get matchmaking as an algorithm: brand chemistry turned into probabilities, then tuned by the market's answer.
Catalog record
| CREATOR | Xiangxiang Dai, Xiaowei Sun, Jinhang Zuo, Xutong Liu, John C.S. Lui |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-22254
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Unified Online-Offline Framework for Co-Branding Campaign Recommendations Xiangxiang Dai, Xiaowei Sun, Jinhang Zuo, Xutong Liu, John C.S. Lui Abstract . Co-branding has become a vital strategy for businesses aiming to expand market reach within recommendation systems. However, identifying effective cross-industry partnerships remains challenging due to resource imbalances, uncertain brand willingness, and ever-changing market conditions. In this paper, we provide the first systematic study of this problem and propose a unified online-offline framework to enable co-branding recommendations. Our approach begins by constructing a bipartite graph linking “initiating” and “target” brands to quantify co-branding probabilities and assess market benefits. During the online learning phase, we dynamically update the graph in response to market feedback, while striking a balance between exploring new collaborations for long-term gains and exploiting established partnerships for immediate benefits. To address the high initial co-branding costs, our framework mitigates redundant exploration, thereby enhancing short-term performance while ensuring sustainable strategic growth. In the offline optimization phase, our framework consolidates the interests of multiple sub-brands under the same parent brand to maximize overall returns, avoid excessive investment in single sub-brands, and reduce unnecessary costs associated with over-prioritizing a single sub-brand. We present a theoretical analysis of our approach, establishing a highly nontrivial sublinear regret bound for online learning in the complex co-branding problem, and enhancing the approximation guarantee for the NP-hard offline budget allocation optimization. Experiments on both synthetic and real-world co-branding datasets demonstrate the practical effectiveness of our framework, with at least 12% improvement. DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation emai; June 03–05, 2018; Woodstock, NY Conference: Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2; August 3–7, 2025; Toronto, ON, Canada Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2 (KDD ’25), August 3–7, 2025, Toronto, ON, Canada DOI: 10.1145/3711896.3736824 ISBN: 979-8-4007-1454-2/2025/08 CCS: Information systems Recommender systems CCS: Theory of computation Online learning algorithms CCS: Theory of computation Approximation algorithms analysis † † cc-license: by KDD Availability Link: The source code of this paper has been made publicly available at https://doi.org/10.5281/zenodo.15532548 . 1. Introduction Co-branding recommendation is a popular strategy that businesses and consumers alike embrace. By leveraging collaborations between two different brands, this recommendation strategy involves co-creating new products or services to maximize user satisfaction and achieve joint market benefits ( Gogri, 2022 ) . For example, the launch of CASETiFY ’s collaborative line of electronic accessories in partnership with Netflix ’s television series ‘‘Squid Game 2’’ strongly resonates with fans of the series, showcasing the success of such brand collaborations 1 1 1 https://www.casetify.com/co-lab/squid-game-2024 . These collaborations not only provide brands with enhanced exposure and increased market gains, but also cater to consumer demand for novelty. According to SocialBeta’s data ( Lite, 2024 ) covering the period from March 2022 to February 2023, approximately 15% of the 4,428 recorded brand marketing cases involved co-branding. Further analysis reveals that this proportion never dropped below 10% in any given month, with co-branding seeing notable spikes during key promotional periods like summer sales and the year-end holiday season (e.g., Christmas and New Year). In this context, co-branding has now become a trend. Despite its potential, a key challenge remains: How can brands identify the most suitable partners from the set of target brands to ensure successful co-branding and maximize market returns? This question remains systematically unresolved and cannot be overlooked. For instance, the collaboration between Heytea and the Jingdezhen China Ceramic Museum resulted in reduced profits, highlighting the importance of selecting the right target brands ( Turan, 2022 ; Helmig et al., 2007 ) . However, this is a challenging task. The co-branding market is fraught with uncertainty. Firstly, overemphasizing certain popular co-branding collaborations can lead to a resource imbalance, i.e., “Matthew Effect” ( Gao et al., 2023 ) , where smaller, long-tail brands are overlooked despite their potential to deliver greater market benefits. The willingness of target brands to participate in collaborations is also uncertain, influenced by multiple factors such as brand positioning and financial commitment from the initiating brand. Additionally, market changes make over-reliance on historical data risky. For instance, the continued use of certain types of co-branding may lead to consumer fatigue or shifting preferences, which may be overlooked by brands overly dependent on past performance. In such cases, collaborations with new brands can disrupt the status quo and introduce a sense of novelty ( Bobadilla et al., 2013 ; Turan, 2022 ) . In such a dynamic online recommendation market, static strategies are insufficient for determining appropriate co-branding partnerships. Instead, a more dynamic approach that continuously adapts to uncertainty is required . This involves addressing the “ exploration-exploitation trade-off ”, where brands must balance the known willingness and market benefits of established partnerships while exploring potential collaborations with new, promising target brands that could deliver better market performance. Recent studies have shown that multi-armed bandit (MAB) methods can effectively address this trade-off challenge in recommendation systems in a lightweight, online manner ( Heyden et al., 2024 ; Li et al., 2025b ; Xia et al., 2023 ; Dai et al., 2024c ) . From a decision-making perspective, co-branding can be framed as a recommendation problem, where the goal is to suggest partnerships that maximize long-term market impact. However, unlike traditional recommendation tasks (e.g., recommending a movie to a user), the costs associated with co-branding decisions are significantly high ( Yang et al., 2024 ; Dai et al., 2024b ; Liu et al., 2024 ) . Implementing a co-branding partnership requires substantial upfront efforts, including extensive market research and creating branded materials, all of which incur significant costs. As a result, while long-term brand objectives may benefit from the online adjustment of co-branding strategies to maximize collaboration gains, exploration must be conducted with caution to mitigate risks—especially in the early stages—to prevent brands from losing confidence in the algorithm before its benefits materialize. In addition to the inherent uncertainty surrounding the willingness or feasibility of co-branding and market gains, brands must also manage budget constraints when planning co-branding initiatives. In practice, the parent brand that initiates co-branding campaigns often oversees “ multiple sub-brands ” (e.g., the Coca-Cola company owns sub-brands like Sprite and Fanta , or Marvel has various superhero intellectual properties (IPs), such as Spider-Man and Iron Man). An illustration of this structure is shown in (a) part of Figure 1 , with further details provided in Section 2 . Due to financial limitations, it may not be feasible for every sub-brand to engage in co-branding with target brands. Consequently, when planning and allocating budgets for a co-branding initiative, the parent brand must adopt a holistic view of the entire brand ecosystem, rather than focusing on a single sub-brand. Focusing solely on co-branding efforts for individual sub-brands could reduce overall collaboration benefits, potentially affecting the profitability of the parent brand as a whole. In such cases, allocating excessive resources to a single sub-brand’s co-branding efforts may lead to diminishing returns, whereas a more cost-effective sub-brand might generate higher overall value for the parent brand. Therefore, it is crucial to assess and align the interests of all sub-brands within the parent brand’s portfolio. Optimizing budget allocation for