A framework to open the black box of the use phase in circular economy life cycle assessments: The case of shell jacket reuse

Life cycle assessments of circular economy measures (CE LCA) of consumer products have been criticized for oversimplifying important aspects of the use phase such as user behavior and rebound effects, limiting our understanding of the environmental performance of circular economy measures. This study tests the usefulness of a framework designed to facilitate accounting for such aspects, by applying the framework to a case study of reuse of shell jackets enabled by “premium secondhand” outdoor stores. Methods for collecting use phase data were user surveys and interviews with store managers. Using the framework on this case study generated several novel insights which are interesting in themselves and as inputs to CE LCA. For instance, secondhand shell jackets have a significantly lower frequency of use during their first use span compared to the second and to shell jackets in the linear reference scenario. This implies that reuse in this case does not function as a mere use extension of otherwise similar use phases as is commonly assumed. The generation of such insights, which hitherto have been lacking in CE LCAs, points to the usefulness of the framework as a tool for opening the “black box” of the use phase in CE LCAs to improve understanding of the environmental performance of circular economy measures.

. Such concerns are connected to aspects of the use phase of products that tend to be oversimplified or overlooked in state-of-the-art CE LCAs, relating to: the products as such, for example, their functionality (Kjaer et al., 2016), the behaviors of users Niero et al., 2021), or the effects CE measures have on products and/or users on a system level (Castro et al., 2022;Zink & Geyer, 2017).
In CE LCA, it can be difficult to define the functionality of products, to ensure that compared alternative products are actually comparable in terms of delivering the same functionality and to account for potential differences between alternatives that may be relevant (Kjaer et al., 2016).
As argued by Kjaer et al. (2016), the functionality of products and whether products are comparable is largely a matter of the requirements of their users. The CE literature has however largely neglected the importance of users (Camacho-Otero et al., 2018;Lazarevic & Valve, 2017) despite that the success of CE depends fundamentally on engaged users (Hobson, 2020;Lazarevic & Valve, 2017). It is the everyday decisions regarding when and how to obtain, use, and rid themselves of products (Selvefors et al., 2019) that makes, or does not make, the CE "go round." In LCA in general (Polizzi di Sorrentino et al., 2016), and in the specific case of CE LCAs , it is customary to model the use phase based on secondary data or pure assumptions. To increase the accuracy of LCA results it is crucial to model the use phase in greater detail, especially when results are sensitive to heterogeneous user behavior (Polizzi di Sorrentino et al., 2016) as they are likely to be in the case of CE measures. A review of CE LCAs however concludes that lack of quality data on product use and user behavior significantly hampers accurate assessment and understanding of the environmental performance of CE measures .
A better understanding of users is also crucial to accounting for the effects CE measures have on products and/or users on a system level (Castro et al., 2022;Polizzi di Sorrentino et al., 2016;Vivanco & van der Voet, 2014). Two important, but commonly overlooked or oversimplified, aspects in this regard are displacement and rebound effects (Castro et al., 2022;Polizzi di Sorrentino et al., 2016;Vivanco & van der Voet, 2014;Zink & Geyer, 2017). Displacement is the extent to which circular products are obtained instead of new ones and thus displace new production. Circular products seldom compete on the same market and attract the same users as new products. For such reasons, it has been argued that circular products do not displace new production to the extent that is commonly assumed in CE LCAs (Zink & Geyer, 2017).
If displacement is imperfect (i.e., when circular products do not displace new ones 1:1) CE measures may merely increase the total supply of products, lower prices, and increase overall consumption and thus generate rebound effects. The concept of rebound effects originates from the context of energy efficiency, but is also relevant and studied in other sustainability contexts (Vivanco & van der Voet, 2014). In the CE context, they refer to instances when "increases in production and consumption efficiency are offset by increases in production and consumption" (Zink & Geyer, 2017). In addition to imperfect displacement, rebound effects of CE measures may be caused by users re-spending income effects which may arise due to circular products being cheaper than new ones (Makov & Vivanco, 2018;Zink & Geyer, 2017). CE LCAs that overestimate displacement rates and fail to account for rebound effects could overestimate potential environmental benefits or overlook instances where CE measures have limited benefits or even increase impacts overall (Zink & Geyer, 2017;Zink et al., 2016). Salient research gaps on rebound effects of CE include: knowledge about behaviors of different user groups; acquiring empirical evidence of different types of rebound effects; and, the influence of rebound effects on the environmental performance of CE measures (Castro et al., 2022).
As a response to these limitations of state-of-the-art CE LCAs and associated research gaps, a framework was developed with the ultimate aim of contributing to better understanding of the environmental performance of CE measures (André & Björklund, 2022). More specifically, the purpose of the framework is to guide CE LCA practitioners in inventorying the use phase more comprehensively than state-of-the-art CE LCAs.
The purpose of this paper is to test and validate the framework. We do this by applying it as a guide for collecting and mapping comprehensive use phase data for a case study of a specific CE measure: reuse of shell jackets for outdoor sports and recreation, enabled by "premium secondhand" outdoor stores in Sweden. The validation of the framework rests upon its ability to help uncover insights about the use phase of this particular CE case concerning aspects that may have significant influence on its environmental impacts. This specific case study is chosen, inter alia, because it is hypothesized to include user groups with distinct characteristics and behaviors, which could be conducive to the validation. The insights gained from this comprehensive mapping are subsequently utilized to inform the goal and scope and lifecycle inventory of a CE LCA, presented in a forthcoming paper, in which the use phase is not "black-boxed." The paper is organized as follows: Section 2 describes key features of the conceptual framework. Section 3 presents background information for the case study and the methods for data collection. Section 4 presents and discusses case study results on selected framework aspects. Section 5 discusses validity and limitations of case study results (Section 5.1) and draws on the results to discuss the usefulness and limitations of the framework (Section 5.2). Section 6 summarizes the main conclusions.

FRAMEWORK FOR MAPPING CE LCA USE PHASE
This section describes the logic of the framework, its components, subcomponents and aspects, and the relations between them. The framework is designed to be applied specifically to consumer products and CE measures that occur in the use phase (André & Björklund, 2022). It was developed with a focus on use extension measures, for example, reuse and repair and is applicable also to use phase efficiency measures, for example, share (Böckin et al., 2020).
Applying the first version of the framework (André & Björklund, 2022) to the case study revealed weaknesses and helped develop it into the current version ( Figure 1) which will be further explained in the following sections. The main changes between the first and current versions are: reorganization into one comprehensive figure to clarify the relations between framework components and aspects mentioned under each component (product and user characteristics, product-user lifecycles, and subcomponents of CE LCA); further theoretical depth, for example, with regard to aspects of user characteristics; and, a more elaborated conceptualization of product functionality. The reader is referred to the publication of the first version (André & Björklund, 2022) and Supporting Information (S1) for more background on the development of the framework, its potential virtues, and applications as well as further motivations for the included components and aspects.
In short, the logic of the framework (Figure 1) can be described as follows: the ultimate goal of facilitating improved CE LCAs (right side) requires more comprehensive mapping of product-user lifecycles (middle). In turn, a comprehensive mapping requires not only establishing what product-user lifecycles look like, but also why. This requires knowledge about product and user characteristics (left side) which interplay to shape product-user lifecycles ( Figure 1). In this section, we will first describe what product-user lifecycles are, what aspects they comprise, and why they are crucial to the results of CE LCAs (2.1). Thereafter, we will explain how product-user lifecycles are shaped by product and user characteristics and describe what aspects they comprise (Section 2.2).

Product-user lifecycles
The term product-user lifecycle is meant to accentuate that a lifecycle studied in a CE LCA is created by the interplay of product and user characteristics. Accordingly, aspects of product-user lifecycles have a product-related side and a user-related side. Product-user lifecycles comprise two subcomponents that are key to the results of CE LCAs: product functionality and product use. Product functionality is key to the results of CE LCAs because functionality, as represented by the functional unit, is key to LCA in general. Product use is key to CE LCA results because it affects how much functionality is delivered to users (use span which refers to the duration in use of each user, lifespan which refers to the sum of all use spans and frequency of use) and how much additional resources are required (auxiliary materials and energy) for using the products. In terms of product functionality, the notion that aspects of product-user lifecycles have a product-related and a user-related side, is reflected by the distinction of states of, and user requirements on, product properties (Van Nes, 2016). As an illustration of how this notion applies to aspects of product use, one can consider the example of lifespans, which are largely determined by technical aspects as well as user requirements.
A list of product properties is located at the intersection between product-user lifecycles and product and user characteristics ( Figure 1). This location signifies that product properties are both product characteristics (which will be described later) and aspects of product functionality. Each product property can be described in terms of state and user requirement. The relation between state and user requirement is referred to as the user valuation of product property. States and user requirements of product properties can be plotted to demonstrate how they develop over each use span, that is, from obtainment to riddance or noted as more general features that are independent of lifecycle stage.
Such plots give insight into product functionality from a broad perspective, which is useful for formulations of functional unit definition, and provide explanations for patterns of product use aspects, for example, use span and frequency of use. The functional unit in LCA is a quantified measure of the function of a product to which environmental impacts are linked (ISO, 2006). In comparative LCAs, for example, CE LCAs, alternative products must deliver equal amounts of function. However, it is not always straightforward to define what the function of a product is. For example, to some users the most important property of a product may be its technical performance, while to others, it may be its appearance. Since new and circular products may perform equally in terms of some product properties but differently in others, it can be important to consider how they compare in terms of a wide range of product properties, for example, through a broad functional unit definition (Goedkoop et al., 1999;Kjaer et al., 2016). A broad functional unit could include what is referred to as both obligatory and positioning properties (Weidema et al., 2004). Obligatory properties are those a product must have to be considered a relevant alternative in comparative LCAs, for example, CE LCAs. Typically, they are related to the principal function of the product and used as a basis for strict functional unit definitions. Positioning properties are market-relevant, but not obligatory, properties that are valued by users (Weidema et al., 2004). For instance, appearance is a property that is often valued by users in addition to the principal function.
The relations between state and user requirement of product properties (i.e., the user valuation of product properties) provide insight into aspects of product use. For instance, if users become dissatisfied with any relevant product property this may lower the frequency of use or prompt a riddance decision (Van Nes, 2016), that is, limit a use or lifespan. Conversely, collecting data on riddance decisions can be used as an indication of how users' valuation of product properties have changed over time. For instance, if the motivation for riddance is that a product does not fulfil performancerelated requirements, this is an indication that either the state of performance has declined or that the user's requirements on performance has increased. Asking users about motivations for riddance is likely to be more feasible and time efficient than asking them how the state and their requirements on all product properties have changed over time. Pace of change of product properƟes (e.g., performance, appearance) Product complexity F I G U R E 1 Framework for comprehensive mapping of the use phase in life cycle assessments of circular economy measures (CE LCAs), adapted and extensively updated from (André & Björklund, 2022). Legend: arrows denote states of, and user requirements on, product properties, which may change over time, for example, from obtainment to riddance. The vertical double-ended arrow denotes the resulting user valuation of product properties, which is the relation between state of, and user requirement on, a product property. Other arrows denote relations between product-user lifecycles and their input to CE LCA. Now that we have described what product-user lifecycles are and what they comprise, we turn to how they can influence aspects of CE LCAs (represented by the arrows from product-user lifecycles to CE LCA in Figure 1). As product functionality and product use determine how much functionality is delivered to users and how much additional resources are required for using products, they shape reference and inventory flows, that is, lifecycle flows.
Comparisons of alternative product-user lifecycles shape and inform assessment of system characteristics such as displacement rates and rebound effects. For instance, the price difference between circular and new products can result in income effects which, in combination with data on users' marginal spending, can be used to assess the re-spending type rebound effect. However, user behavior, and by extension rebound effects, is not only determined by price and income, and users are not merely seeking to maximize utility but rather strive for subjective well-being and quality of life (Hofstetter et al., 2005). Hence, other determinants of user behavior, that may be referred to as socio-psychological, such as environmental attitudes, social norms, and social status (see Section 2.3) are also relevant to the assessment of rebound effects (Castro et al., 2022;Hofstetter et al., 2005). However, it is difficult to establish the causal mechanisms of rebound effects (users themselves may not be able to explain their behavior) (van den Bergh, 2010) and mechanisms may be synergetic. In order to point to relevant data, the framework distinguishes between economic mechanisms, shaped and informed by price data, for example, re-spending, and socio-psychological mechanisms, for example, moral licensing (Dütschke et al., 2018).
The latter are shaped, and can thus be informed, by data on user characteristics, use span, riddance motivations, and product properties related to, for example, environment, social affiliation, and enjoyment.
Data on product functionality in terms of obligatory and positioning properties (Weidema et al., 2004) informs methodological choices such as functional unit definition. In addition, the distinction between obligatory and positioning properties is relevant to displacement rates. There are two main approaches to modeling displacement: technical and market equivalency (Vadenbo et al., 2017). Technical equivalency is based on the relation between circular and new products in terms of quantifiable measures relating to principal functions, often obligatory properties. The market equivalency approach is rooted in the notion that obtainment of circular products may not displace production of new ones, even if they are technically equivalent, since they may not compete on the same market (Vadenbo et al., 2017;Zink et al., 2016). Thereby, the market equivalency approach can be described as being based on the relation between circular and new products in terms of all market-relevant properties, that is, both obligatory and positioning properties. Displacement rates are commonly calculated from data on obtainment motivations that serve to discern to what degree obtainments were motivated by a real need for the product or, for example, a lower price (Farrant et al., 2010;Stevenson & Gmitrowicz, 2012).
Because such data includes other product properties than technical performance of principal functions, for example, price, deriving displacement rates from obtainment motivations can be seen as a form of market approach.

Product and user characteristics
A deep understanding of the environmental performance of CE requires not only mapping product-user lifecycles as outlined above. It requires knowledge of the underlying reasons behind product-user lifecycle patterns, that is, the determinants that shape product-user lifecycles. These determinants are categorized into product characteristics, user characteristics, and contextual factors 1 (Van Nes, 2016) ( Figure 1). Product characteristics compose both product properties, for example, performance and appearance, and other, more overarching, product characteristics, for example, the pace of change of product properties (Böckin et al., 2020). The list of product properties included in the framework is inspired by works that have aimed to understand the user experience of engaging with circular products (Bigliardi et al., 2020;Laitala & Klepp, 2018;Schallehn et al., 2019). This list is not exhaustive but is believed to provide a good starting point for many CE LCAs. Data on how users value new and circular products in terms of this range of product properties can help explain why some users prefer new over circular products and vice versa. As examples of the relevance of overarching product characteristics it can be mentioned that the speed of technical innovation influences the performance of new products relative to circular ones and that product complexity may influence the feasibility of CE measures, for example., hinder repair (Böckin et al., 2020).
User characteristics refer to aspects that influence environmental behavior: attitudes, personal capabilities, and habits (Stern, 2000;Testa et al., 2020;Van Nes, 2016). Attitudes influence user requirements and valuation of product properties. Personal capabilities refer to such things as financial resources, which influence income effects and rebound effects, or knowledge and skills which could refer to users' capacity to repair products, and thus, extend use spans.
In addition, interplays between product and user characteristics are influenced by contextual factors such as social norms, legislation, and life events (Madon, 2022;Stern, 2000;Van Nes, 2016).

METHODS
In this section, we motivate and provide background information for the case study (Section 3.1), describe the methods used for data collection (Section 3.2), and motivate the selection of framework aspects to focus on in this paper (Section 3.3).

Case study
The case study focuses on two "premium secondhand stores" located in Sweden that sell outdoor clothing and equipment for outdoor sports and recreational activities: Tracks Recycle, founded in 2017 and located in Gothenburg and Re-Adventure, founded in 2020 and located in Stockholm. A reference scenario was created based on user surveys at five conventional stores for outdoor clothing and equipment that predominantly sell new products, located in Gothenburg.
This case study was chosen for a number of reasons. Importantly, it was hypothesized that the outdoor equipment market would comprise discernable user groups with distinct characteristics and behaviors. The study of such distinct characteristics and behaviors was hypothesized to be conducive to testing aspects of the framework. In addition, CE measures for outdoor equipment play out locally in Sweden to large extent, in contrast to many other product categories that have largely international geographical scopes. This allows for inventorying the entire product-user lifecycles (or at least the most relevant parts). Another practical reason is that the stores are physical (with online sales as well) which facilitates acquiring knowledge about users: both directly through surveys and indirectly through interviews with store managers. Shell jackets were chosen as a representative product of the secondhand stores in dialogue with the managers due to high turnover and fairly even demand all year round.
Both secondhand stores have similar business models and markets. Profits of resold products are shared with the previous users who hand in the products for resale. The "premium" label refers to their primary focus on high-quality products from reputable brands. Prices are set with consideration to mainly firsthand retail price and product condition. The purpose of studying both stores was primarily to generate more robust findings by increasing the number of survey respondents and observations by store managers.
The product-user lifecycles that include reuse through a resale at the secondhand stores are referred to as the "2U" lifecycle, denoting that shell jackets are used by (at least) two principal users during their lifetime. The reference scenario is referred to as the "1U" lifecycle, denoting that shell jackets are used by one principal user during their product lifetime. In the 2U lifecycle, the first users, who sell products at the secondhand stores are referred to as "2U1-users" and the second users, who buy products there are referred to as "2U2-users." Users of the 1U lifecycle are referred to as "1U-users."

Data collection
Data for the inventory of the use phase was collected through interviews and surveys. The combination of the two methods enables data triangulation (Verschuren et al., 2010) which strengthens the validity of the empirical findings. Interview and survey questions were developed based on the framework, literature on framework aspects that were deemed relevant for the case study, for example, displacement (Farrant et al., 2010;Stevenson & Gmitrowicz, 2012) and consumer research (Klepp & Bjerck, 2012;Laitala & Klepp, 2018).
First, semi-structured interviews were conducted with store managers. The main aim was to provide qualitative data on user characteristics of discernable user groups, for example, attitudes and personal capabilities (available in Supporting Information S1). A secondary aim was to create an initial understanding of product-user lifecycles, contributing to the formulation of survey questions. The survey was intended to provide quantitative data to be used as basis for the CE LCA. For this reason, it included mainly closed questions. Open follow-up questions were also included to allow respondents to elaborate. The surveys were accessed through QR codes placed in the physical stores and through links distributed through the stores' communication channels. Since the use spans of 2U1-users had ended, their responses were retrospective. The use spans of 2U2-users and 1U-users were on-going, making their responses partly prospective. More information about the surveys and potential differences between retrospective and prospective responses are available in Supporting Information S1 and S2 as well as Section 5. The total number of survey respondents amounted to 71 users, comprised of 30 2U1-users, 21 2U2-users and 20 1U-users. In Section 5.1, we discuss what this limited sample entails for the validity of results.

Selection of framework aspects
This paper presents a comprehensive mapping of product-user lifecycles of the case study (middle column in Figure 1), supported by data on product and user characteristics (left column in Figure 1) while the utilization of the mapping in the CE LCA (right column in Figure 1) of the case study is presented in a forthcoming paper. Although the comprehensive mapping is ultimately intended to inform the CE LCA, it provides insights that are interesting as stand-alone results, in terms of revealing novel knowledge about aspects that may be influential for the environmental performance.
Framework aspects related to these specific insights are the focus of this paper: product functionality and the aspects of product use: use span, lifespan, and frequency of use (Supporting Information S1, Figure 3).

RESULTS AND ANALYSIS
The mapping of selected framework aspects is organized into the following sections: 4.  Figure 2 illustrates how users value product properties when buying shell jackets, that is, user requirements. The results are overall similar for the three user groups. In general, all user groups rank product properties in the same order of importance. All user groups consistently value performance and quality highly. Hence, these product properties can be categorized as obligatory.

Product functionality
Below, we discuss some small differences between the user groups. Although these differences are not statistically significant, they are nonetheless logically reasonable and reinforced in combination with interview data (Figure 2).
-Price appears to be slightly less important to 2U1-users than to the other user groups. This is aligned with interview data indicating this is a group of higher level of income than most 2U2-users (A. Mangs Bergmark, personal communication, August 24, 2021).
-Environment appears to be less important to 1U-users than 2U-users. A plausible explanation is that a higher valuation of environment could be a reason for 2U-users to engage with the secondhand stores.
-Appearance appears to be most important to 1-and 2U1-users and least important to 2U2-users. This is reasonable considering that new jackets are likely to be assessed as having a higher state in terms of appearance than secondhand jackets.
-Social affiliation appears to be most important to 2U2-users. This could be related to that some of them consider buying products secondhand to be part of their lifestyle (K. Olivensjö, personal communication, July 26, 2021).
-Convenience appears to be most important to 1U-users, which is reasonable given that conventional stores tend to have, for example, larger selection of models and sizes.
-Enjoyment appears to be most important to 2U-users. For 2U1-users, data presented in the remainder of the results section suggest the valuation of enjoyment could be linked to a novelty-seeking behavior (i.e., the part of our description of enjoyment saying "it should. . . give pleasure of novelty" (Supporting Information S2)). For 2U2-users, the valuation of enjoyment could rather be related to the part of our description of enjoyment saying "it should feel good to purchase" (Supporting Information S2) and reflect their inclination to buy secondhand, assuming secondhand consumption makes them feel better about a purchase than if buying new products.
Figure 3 displays how the product condition, as assessed by users, declines over the course of the use phase. 2 It is based on users' assessment of shell jacket condition at points of obtainment (2U1, 2U2, and 1U), riddance (2U1) and expected condition at riddance (2U2 and 1U) and assuming a linear decline between these points (Supporting Information S1 and S2). It is worth noting that 2U1-users decide to sell their shell jackets despite them being in a condition between "excellent" and "good." At end of use, shell jackets are expected to be in a condition between "decent" and "poor" in both alternatives. Overall, the lifecycles of both alternatives are fairly similar in terms of product condition and lifetime.

Product use
Figure 4 displays product use data in terms of frequency of use, use span, and lifespan. 1U-users report that they use their shell jackets about three times more frequently, and 2U2-users about twice as frequently compared to 2U1-users. This results in the increase in frequency of use at the point of secondhand resale, after 3−4 years in the 2U lifecycle. The second use span of the 2U lifecycle is quite similar to the 1U lifecycle: in addition to a comparable frequency of use, shell jackets are reported to be used for 7−8 years. The main difference between the alternatives is that the main use of shell jackets is delayed for 3−4 years in the 2U lifecycle, that is, the duration of the first use span. In total, shell jackets in the 1U lifecycle are used more than shell jackets in the 2U lifecycle (approximately 800 and 625, respectively).

F I G U R E 5
Motivations for riddance among 2U1-users. The underlying data for this figure can be found in Supporting Information S3. Figure 5 shows the motivations for riddance among 2U1-users. For the majority, the motivation is not that shell jackets do not longer fulfil their functional or aesthetic requirements. Neither is it because they lack space, nor need, or want money. Instead, the most relevant reason is they "have another one or plan to buy another one." This prompts the question what motivates the users who "have [already acquired] another one or plan to buy another" despite the previous one having fulfilled their (functional and aesthetic) user requirements. 3 The collected data suggests the most plausible explanation is that these users are motivated by a desire for novelty. Some 2U1-users are "regulars" who return often to sell "new, nice and expensive gear" in rather unused condition when new "new, nice and expensive gear" is released on the market (A. Mangs Bergmark, personal communication, August 24, 2021) (Supporting Information S1). The prevalence of such a novelty-driven user behavior is supported by an employee at Patagonia (Jenni Swadding, personal communication, May 27, 2021) who reports that releases of products in new colors often result in noticeable consumption. 2U1-users that fit the description of these regulars 4 consider the product property of "enjoyment" to be almost twice as important as the average 2U1-user when buying shell jackets (Supporting Information S1). Altogether, this suggests that there indeed seems to be a desire for novelty that motivates the riddance decision for many 2U1-users.

Motivations for riddance and obtainment
Further, the user behavior of the "regulars" suggests that they are more interested in buying and owning expensive gear than using it (A. Mangs Bergmark, personal communication, August 24, 2021). This is one explanation for the significantly lower frequency of use among 2U1-users compared to other users ( Figure 4). The other explanation is that users with low frequency of use reported that the shell jackets did not fit them or did not perform as expected (Supporting Information S1). In other words, the low frequency of use among 2U1-users is caused by an underlying discontentment that also motivates the riddance decision.
The motivations for obtainment among 2U2-users ( Figure 6) are investigated to inform the analysis of displacement. Agreement with the first and third statements correlate with low displacement while agreement with the other statements correlate with high displacement. Price seems to have influenced most users to varying degrees. There is low agreement with the intention to use the shell jacket as a backup. Most purchases are motivated by "a need for a shell jacket" and many visited the secondhand stores with the specific intention of buying a shell jacket. However, there are also many who did not visit the stores with this specific intention. Thus, some obtainments seem to be based on impulse. Overall, secondhand purchases are indicated to displace new production of shell jackets to a considerable degree (quantitative displacement rates are quantified in the forthcoming paper on the CE LCA of the case study), but definitely not at a 1:1 ratio as is commonly assumed (Zink & Geyer, 2017).
Around 70% of 1U-users prefer buying new over secondhand (Supporting Information S1). The motivations for this preference are well aligned with their user requirements when buying shell jackets. For example, users who report that convenience is very important when buying a shell jacket typically report that convenience is very important to their preference for new over secondhand. The main motivations for preferring new over secondhand jackets are that new ones last longer, are more convenient to purchase, can be fully trusted and that there are no knowledge gaps on how they have been previously used or cared for. One property, enjoyment, is noticeably more important for the preference for new over secondhand shell jackets than as a user requirement at obtainment in general. In other words, it seems influential to the 1U-users preference for new, over secondhand shell jackets, that the purchase of new shell jackets give them more enjoyment than the purchase of secondhand shell jackets (Supporting Information S1).

F I G U R E 6
Motivations for obtainment among 2U2-users. The underlying data for this figure can be found in Supporting Information S3.

F I G U R E 7
Expected motivations for riddance at end of use for 1U-users. The underlying data for this figure can be found in Supporting Information S3.
1U-users expect that their future motivations for riddance will be related to performance, quality (that it breaks), and fit (Figure 7). The alternative "that I have or plan to buy another one" is much less relevant compared to 2U1-users, but still quite relevant.

Income effects and marginal spending
The inventory of how price changes over the lifecycle informs the analysis of potential income effects of secondhand sales that could be re-spent to generate rebound effects (quantified in the forthcoming paper on the CE LCA of the case study). 2U1-users receive on average about 900 SEK (around 90 EUR) when selling their shell jackets (50% are retained by the stores). On the other hand, the practice of 2U1-users, that is, buying new and selling at secondhand, costs on average about 700 SEK per year of shell jacket ownership (["original price" − "secondhand price"*0,5]/"use span"). With this in mind, it seems questionable to consider money from the resales as an income effect. Nonetheless, money from the resales may very well be perceived as re-spendable income by these users.
According to the interviews, buying new shell jackets of the same type at full price does not seem to be a real option for many of the 2U2-users (K. Olivensjö, personal communication, July 26 2021;A. Mangs Bergmark, personal communication, August 24, 2021). The majority of 2U2-users have low income and are price sensitive, for example, students, who would not prioritize or could not afford buying jackets from the same high-quality reputable brands as they are able to at the secondhand stores. Therefore, many of the secondhand purchases seem to result in upgrading of quality rather than saving money, that is, resulting in no income effects. However, there are also high-income 2U2-users, for whom income effects may occur (K. Olivensjö, personal communication, July 26 2021;A. Mangs Bergmark, personal communication, August 24, 2021).
A condition for high likelihood of rebound effects is if income effects are spent on more impactful consumption than the original type of consumption (Makov & Vivanco, 2018;Zink & Geyer, 2017). In this case, however, marginal spending mostly involves the same type of consumption, that is, outdoor clothing and equipment (Supporting Information S1). About half of the 2U1-users (14/30) report they will spend the money from the resale (Supporting Information S1, Figure 10), most (11/14) on outdoor equipment, in particular, shell jackets (9/14). In contrast to 2U1-users, most 2U2-users will do "nothing in particular" with potential income effects. These answers are well aligned with the previously mentioned price sensitivity and low income of many of these users.

DISCUSSION
The results of this case study are primarily interesting in terms of their validation of the framework, which is discussed in Section 5.2. Some are however empirically interesting, why their validity first needs to be discussed with respect to quantity and quality of data as well as case study design (Section 5.1).

Validity and limitations of case study results
The number of survey respondents was limited by the number of users selling or buying shell jackets in the studied stores. Having more respondents would likely have changed the exact figures on, for example, user behavior with regard to product use. Because of this, these results are valid as indications that certain user behaviors exist, but not as representations of how common they are. Despite not being statistically significant, results based on such limited samples of use phase heterogeneity are still valuable in LCA for generating more valid interpretations than if the data were not collected (Daae & Boks, 2015).
A potential problem with surveys is that respondents may seek to present themselves in a socially admirable light, potentially resulting in social desirability bias (Stocké & Hunkler, 2016). In this case, respondents could be reluctant to admit having a high valuation of product properties such as "price," "social affiliation," and "enjoyment" and, conversely, exaggerate their valuation of properties such as "environment." While survey results are based on stated preference, interviews with managers could be seen as rather based on revealed preference, since managers' impressions of users are rather based on revealed user behavior. In this regard, the interviews complement the user surveys by providing qualitative insights exempt from potential social desirability bias.
A social desirability bias could also influence responses on use and lifetimes, especially prospective responses (i.e., 1 and 2U2-users). Environmental results of CE LCAs are often sensitive to use and lifetime, why this potential social desirability bias is important to consider in the CE LCA of the case study. The lifespans reported by users (1U: 7−8 years; 2U1: 3−4 years + 2U2: 5−6 years = 9--10 years) are slightly longer compared to another survey (European Outdoor Group, 2015) where users' average intended use span was 5−6 years and the perceived lifespan among brands of their best-selling shell jacket was 7−8 years. The indication that the lifespan of the 2U lifecycle is longer than brands' expected lifespan could still be valid considering the limited frequency of use during the first use span, resulting in limited functionality decline. A problem with counterfactual reference scenarios is that they are hypothetical by nature. In this case study, rather than taking a counterfactual approach, we compare two real-world product-user lifecycles, one including resale at secondhand stores and one that does not. However, this realworld comparison also implies that, to some extent, we are essentially comparing lifecycles comprising users who are, on average, content with their shell jackets (1U lifecycle) with lifecycles where first users are not entirely content with their shell jackets (2U lifecycle). This can be seen as a form of selection bias which begs the question: if it is a comparison between average-content and somewhat discontent users, does it provide insights into the environmental performance of CE? We argue that it does. Circular lifecycles are circular precisely because a user in its lifecycle has been somewhat discontent with a product and a subsequent user has found use in it. With this in mind, the short first use span and the low frequency of use of 2U1-users should not be surprising. Yet, it is a novel and interesting result. In future research, it could be interesting to investigate to what extent this pattern applies to other products.

Usefulness and limitations of the framework
The proposed framework addresses the need for better understanding of what shapes consumption patterns (Hofstetter et al., 2005) (c.f., product and user characteristics) and for modeling more realistic and accurate scenarios (c.f., product-user lifecycles) in LCA (Polizzi di Sorrentino et al., 2016), specifically CE LCA Niero et al., 2021). It outlines challenging aspects that are known to be influential for the results of CE LCAs and what type of data could be conducive to addressing them. It does not however specify how such data is to be collected, for example, surveys (Daae & Boks, 2015;Polizzi di Sorrentino et al., 2016), practice theory methodologies (Niero et al., 2021), or direct measurements, for example, via sensors (Polizzi di Sorrentino et al., 2016)) since the best option will differ depending on the goal, scope and resources available. A key contribution of the framework is rather that it provides a comprehensive and coherent structure (drawing on CE frameworks and knowledge from a variety of academic fields) for collecting, organizing, and analyzing such data (André & Björklund, 2022). In future research, the framework could be further developed by application to other cases studies, for example, use phase efficiency measures, for example, sharing (Böckin et al., 2020).
Application to other case studies could also help develop theory on when and why CE measures are likely to be successful in terms of reducing environmental impact.
Given the discussed uncertainty of the case study results, for example, concerning how common specific user behavior patterns are, it may be questioned to what extent the framework is useful as a means for improved inventory of the use phase in CE LCA. The uncertainty of the case study results however pertains to the quantity and quality of data of the case study rather than the framework. Moreover, it needs to be considered what the alternative to collecting this type of uncertain qualitative data is. In state-of-the-art CE LCAs, the alternative has been to simply make assumptions without the support of data . In line with the view of Daae and Boks (2015), the usefulness of the framework and the collection of, albeit uncertain, primary data is that it allows for making better assumptions than without it. As an example, the short first use span and low frequency of use therein could hardly have been assumed without reference to supporting data. The importance of such data for the results of CE LCAs cannot be overstated. The lack of primary data on the use phase of most CE LCAs is probably one reason why this, somehow, reasonable and unsurprising pattern can still be a novel finding.
Based on the short first use span of the 2U lifecycle, the riddance motivation of product replacement and the user characteristics of the user group referred to as "regulars," there is evidence to suggest that the existence of the secondhand market could intensify (new) consumption. In the survey results, the regulars have noticeably higher requirements on appearance and, especially, enjoyment compared to other 2U1-users. This is aligned with the interview data where they are described as people who want "new, nice and expensive gear" more for the sake of status than for using it (A. Mangs Bergmark, personal communication, August 24, 2021). The potential of secondhand markets to reduce costs of ownership and thus intensify (new) consumption for "fashion-conscious" first users was discussed decades ago (Fox, 1957). It has also been discussed as a hypothesis in the more recent CE literature (Cooper & Gutowski, 2017;van Loon et al., 2021) but we are not aware of any empirical support for it prior to this study. This study does not however allow for asserting that these users, in fact, shorten their use spans due to the possibility of secondhand resale. That would require comparing the 2U lifecycle to a counterfactual scenario in which these secondhand stores did not exist, or other forms of research. The study does however establish that the regulars have the required "fashion-conscious" characteristics (Fox, 1957) that could be especially prone to realize this rebound effect, that they receive a financial reimbursement to facilitate it and that the first use span of the 2U lifecycle is less than half compared to 1U-users, that is, those who also buy new jackets but keep them for their entire lifespan. The empirical evidence for this potential rebound effect both points to the usefulness of the proposed framework but also to a limitation of it.

CONCLUSIONS
A framework intended to deepen our understanding of the environmental performance of CE measures for consumer products has been tested and validated on a case study of reuse of shell jackets. The framework facilitates inventorying aspects of the use phase that are commonly overlooked but, in many cases, imperative to CE LCA results. The study demonstrates the usefulness of the framework by generating novel insights about aspects of the use phase of shell jackets that may be influential to the environmental performance of reuse, for instance: -Lifecycles involving secondhand (2U) are quite similar to those that do not (1U) with the exception that most of the active use is delayed for the duration of the first use span (Figures 3 and 4). In fact, shell jackets in lifecycles involving secondhand are used less than shell jackets in lifecycles that do not involve secondhand.
-In lifecycles involving secondhand, the short use span of first users, their main motivation for riddance (product replacement, seemingly driven by a desire for novelty) and their user characteristics provide empirical support for the hypothesized rebound effect (Cooper & Gutowski, 2017;Fox, 1957;van Loon et al., 2021) that secondhand markets could intensify new consumption for some users.
These insights are examples of empirical knowledge that so far has been lacking in CE LCAs. In the secondhand lifecycle, first users' low frequency of use and their decision to sell shell jackets after a short use span appear as two symptoms of an underlying discontentment which is typically rooted in desire for novelty or fitting problems. This pattern should not be surprising. Yet, it is a novel and interesting result, uncovered through the use of the proposed framework. Thereby, the framework has the potential to help practitioners build a richer knowledge of the use phase of consumer products, which is highly needed in CE LCAs in order to advance our understanding of the environmental performance of CE measures.