**5. Discussion**

This section discusses the implications of some of the exposed social housing processes in concert with the possibilities of mass customization.

A vital consideration refers to how public authorities view social housing production. Through legislation and practices, it is clear that the initial housing unit is seen as the final product. All the legislation regarding the production of social housing units refers to the process as finished once the construction of the unit is complete. Post-occupancy social work is the only thing that goes beyond the finalized construction of the initial unit and it does not consider further construction. In contrast, for the families, the initial house unit is seen as a starting point from which they continue to build. As the neighborhood evolves, the self-designed and self-built additions to the units often bring unintended negative consequences along with the intended benefits. The legislation regarding technical assistance in housing design is broad enough that it could include social housing developments. However, it is not seen in practice as applicable in these cases. Pre-occupancy construction, post-occupancy renovation, and technical assistance are organized and seen as completely separate processes. Thus, one happens without any consideration of the other. However, when considering the limits of the initial production of social housing and the needs of the families over time, it becomes clear that for social housing developments, it should all be seen as part of one continuous process, including for funding purposes. Broadening the context considered for the use of funds can lead to a more e fficient allocation of those funds [37].

This is particularly relevant when considering mass customization for this context. Previous studies have indicated that small changes in the current process of provision of social housing could create opportunities for mass customization of the initial units [14]. However, by including the post-occupancy processes in the analysis of potential mass customization possibilities and benefits for this context, it is possible to indicate that customizing the initial unit is not the most sustainable option as it may not have lasting results in terms of satisfying the families' needs and avoiding problematic situations over time.

An important consideration is that mass customizer companies usually profit by attracting more customers with customized products, gaining their fidelity, or taking advantage of their willingness to pay a premium for a custom product [2]. However, these options are not available in this social housing context. The amount the company can receive per unit is capped and the families do not have a choice of developer—they are assigned to a unit by the city. Stakeholders from di fferent spheres of governmen<sup>t</sup> are more interested in a higher number of units built than higher quality in design. This appears in all publicity, and even o fficial program websites state that priority will be given to proposals that reach a higher number of families [38]. Therefore, the companies still would not be motivated to mass customize, even if the added e ffort and cost were only marginally higher.

Regulations around the numbers and types of rooms that the units must have, combined with the restricted floor area to keep the costs low, mean that there is very little that can be customized in terms of the families' space needs. Although these regulations are necessary to establish a minimum standard, they can also be a barrier to mass customization. Furthermore, most changes made to the housing units post-occupancy add area to the house. This shows that in most cases, a large part of the reason for the unit not satisfying the families' needs refers to it being too small. This aspect would still be present even if the initial units were customized. On a di fferent approach, the initial area of the housing units could be increased, making it more meaningful to customize them. However, this would require significant changes in the current process of provision, especially increasing funding. Therefore, this is not a feasible option. Investing in flexibility could be a way to allow for the families to satisfy their needs after occupancy better, while still building all initial units equal within current regulations. However, the way the families currently change and expand their units also leads to problems for themselves, the neighborhood, and the city. Thus, providing assistance for the families in this renovation process is necessary. It is also important to highlight that the needs of the families change over time. Thus, having a strategy to customize the initial unit and not addressing this as a process over time could still lead to the same problems, especially with the expansions.

This study argues that it would be more feasible and have a higher potential for bringing significant benefits over time—not only for the families but also to the city and other stakeholders—if a mass customization strategy were applied with post-occupancy di fferentiation. This way, the initial units built could still be small and equal within the development, not requiring significant changes to current social housing programs, their associated policies, and the capabilities of the stakeholders. Post-occupancy, the families would use a mass customization configurator, a co-design system, to interact with and visualize the design of expansions for their homes. Such configurators have the capacity to validate the customer's solutions, including in cases that involve spatial aspects in housing design, as shown by previous authors [7,8]. Only solutions within the legal parameters for the city would be validated. The design validated by the configurator could receive automatic approval by the city. This would significantly benefit the families and the city by avoiding problematic changes to the units. The families would have the opportunity to manipulate and visualize di fferent design solutions and receive feedback before engaging in construction.

Providing this assistance to the families to have validated design solutions could also serve as a justification for cities to take more severe action, such as fines and demolitions, towards problematic construction situations. The interviews done for this research indicate that there is currently a perception that any actions taken in this regard would be hurting an already vulnerable population and, thus, that such actions are not socially acceptable. This perception is intensified by the fact that most cities do not o ffer an alternative to the current practices of self-design and construction, making it morally and politically challenging to take serious action against such illegal construction. However, if processes such as those proposed in this study are in place and design assistance is available to the families, then it would be a choice of the family to ignore such help and build problematic solutions. This could remove some of the perceived barriers around taking action to correct problematic situations; for example, demolishing construction that encroaches on public space. In other words, if the city is providing the conditions for families to build legally and these families still choose to build illegal solutions, then it could become socially and politically acceptable for the city to fine or demolish such problematic solutions.

To allow post-occupancy di fferentiation as easily and as much as possible, increased adaptability of the initial unit is essential. Most cities do not have the resources to critically evaluate and sugges<sup>t</sup> design changes during the approval process of new social housing developments, with city authorities limited to checking for compliance with legislation. One of the results of this limitation is that most projects merely achieve the minimum required by legislation. In terms of adaptability, the regulation has one sentence that states that the housing unit shall be designed to enable its future expansion without loss to the lighting conditions and natural ventilation of the existing rooms [26]. Given that this regulation provides little guidance, it is usually fulfilled by showing that one room can be added. Often this means that the unit was designed to allow only one room, in a specific place, to be added. This is not enough to satisfy the amount and variety of changes the families need.

In contrast, several researchers have developed extensive adaptability guidelines specifically for social housing. However, most of such guidelines depend on subjective design judgment, making it di fficult to incorporate in regulation and to check for compliance. Although changing the way projects are approved in cities to include their design interests and feedback would be ideal, it is currently not feasible for most cities; thus, providing such thoughtful feedback is outside their capability. An option that could help include more adaptability in design would be to include guidelines in the regulation of the program in a way that would make compliance to them easy to check. One way this could be done would be to use the adaptability guidelines developed in previous studies, for example, by Brandão [18], but translating them as much as possible into quantifiable acceptable ranges that are easy to demonstrate. For the project to be approved, the developer would then have to demonstrate compliance to at least three of the guidelines in addition to the one already in the regulation. Table 1 shows a selection of guidelines that would be feasible to implement within the regulations of the MCMV program and could be checked easily within the current process of approval. The table also shows possible wording for the regulation, such that it would be easy to demonstrate and check for within the approval process. Figure 6 shows an example floor plan inserted in the existing lots of the Anglo neighborhood that was developed in this study following such guidelines. The walls with pipes and structural columns are shown in gray, and the houses are meant for semi-detached typology. Figures 7 and 8 demonstrate some of the expansions possible with this design, which significantly increase the area of the house, demonstrating the design's compliance to the guidelines in Table 1.

This study proposes that the main agen<sup>t</sup> of the mass customization strategy, in this case, should be the local authorities and not a company as in most mass customization cases. This would allow the city to have more power over the parameters to be validated within the configurator and a building system to be used, also making automatic approval feasible. Furthermore, local authorities are more motivated to keep in mind the best interests of the families and the city and could also incorporate the use of the configurator in post-occupancy social assistance within its current format.

To maximize the potential of such a mass customization approach to bring benefits over time, it is essential to consider the families' current post-occupancy processes. For example, distributing the costs of construction over time and being able to self-build to save even more money are important aspects to be maintained. This has direct implications regarding construction for the mass-customized product. For example, pre-fabricating entire rooms to be combined on-site would not be a feasible solution. Pre-fabricating panels that can be combined, allowing families to build at their own pace, as well as using materials that are familiar to them, would be more feasible. Having the local authority as the main agen<sup>t</sup> would also allow several construction companies to be linked to the strategy to fabricate the necessary components.


#### **Table 1.** Adaptability guidelines.

**Figure 6.** Floor plan of an example unit considering the guidelines for adaptability.

**Figure 7.** Floor plan of an example unit with expansions, adding several rooms without the need to destroy walls with pipes from the original unit.

**Figure 8.** Section of two semi-detached example units. The unit on the right shows an expansion along the slope of the roof maintaining the same angle.

Although this study refers to a specific context, applying mass customization with post-occupancy differentiation could be valuable in many other housing contexts. The aspect of the initial unit not being seen as the final product could be easily extended to other housing contexts. In some cases, it may even be easier to overcome some of the challenges faced in the context considered for this study. In North America, for example, lighter building systems, such as wood-framing, are widely accepted. Furthermore, pre-fabrication within those systems is more widely available. As in some examples shown by Smith [39], a company may seek local pre-fabrication options close to each of the sites where they build to keep the costs low. This could be an opportunity for companies to provide an initial small affordable unit combined with a mass customization system of co-design and prefabricated parts or panels of the same building system that the user continues to engage (buy from the company in this case) over time at their own pace.

Aspects of dimensional or geometric co-design in housing are also relevant to discuss. It has been shown that an aspect that can add value to mass customized products refers to the satisfaction of the customer in perceiving themselves as the creator of the product, the "I designed it myself effect" [40]. However, design complexity and substantial investment in the product have shed doubt on whether this factor can be considered in housing. One of the reasons for dimensional mass customization of housing not being more widely available refers to the customers not having enough confidence or knowledge

to take responsibility for the design of their homes; thus, it refers to social and cultural reasons and not technological limitations [5,41]. Many people would prefer to buy the house they perceived as professionally designed, even if they had the option of co-designing. Counteracting this is the need imposed by the context within the lowest income social housing scenario. Many post-occupancy studies [25,32,42] in this context have demonstrated the users' willingness to design their own homes, given that the alternative is not building. Thus, with the use of a co-design system, the resulting designs may still be perceived as more 'professionally designed' than the alternative.

Given the contextual conditions, these social housing neighborhoods could be the ideal place to begin the adoption of such geometric co-design processes, given the reduced emphasis of those social and cultural factors considered as significant challenges to the broader adoption of geometric mass customization in housing. Seeing such processes and their results within this context could be encouraging for developers to adopt geometric co-design in other contexts. Furthermore, it could also be encouraging for the users in overcoming some of their insecurities with what they may perceive as self-design.
