The impact of detergent performance on sustainable consumer laundry behavior: a socio-technical challenge

2.1 Socio-technology interplay and laundry sustainability

Socio-technical systems involve an interplay between a variety of factors and participants including manufacturers, technology, regulators, consumer practices, cultural norms, and infrastructure [28]. Shove [29] presents the case that laundry, while seemingly a routine household task, has all the ingredients of a true socio-technical system (Figure 1). Change in socio-technical systems is difficult because of the need to change not only technology but also the behavior of all the stakeholders in tandem [30]. Transitions in socio-technical systems are often characterized as co-evolutionary because of feedback loops between the different subsystems which modify outcomes and direction [31]. For example, Shove [29] describes the modern washing machine program design as a result of an interplay between clothing fashion designers, synthetic fabrics, environmental concerns, cleanliness standards of excellence, and technology limitations.

Figure 1:

Socio-technical system for laundry. Adapted from [32].

2.2 Laundry performance technology factors

The functional objective of clothes-washing is to remove stains and odors so that clothes look and smell clean [22]. Consumers typically judge this by visually examining clothing for stains and brightness, wrinkles, residues, and presence of a fresh, clean scent [33]. The Sinner cycle, first described by Herbert Sinner in 1959, outlines the four independent factors affecting cleaning performance: chemical action, temperature, mechanical action, and time [34]. Lambert et al. [35] and Lasic et al. [36] examined the influence of detergents, washing temperature, washing time, and load on the washing performance in a systematic way. The results clearly showed an increase in washing performance values by increasing detergency, increasing nominal washing temperature, increasing washing time, and decreasing load sizes. Importantly, these levers are not equal in resource use efficiency creating the opportunity for innovation that maintains washing performance while at the same time reducing environmental impact [37]. Indeed, washing machine eco-program cycles take advantage of this by trading off energy intensive hot water for lower temperatures but longer agitation times [38].

Detergents are frequently in the sights of regulators for lowering environmental impact due to their inherent chemical nature [23, 39]. Modern laundry detergents are a complex mix of surfactants, enzymes, builders, optical brighteners, foam regulators, bleaching agents and other minor additives [40]. Most recently, the addition of enzyme ingredients, which perform equally well at lower temperatures, has increased the cleaning performance in cold-water washing, obviating the need for higher water temperatures [25]. Laitala and Jensen [41] studied the washing effects of eight laundry detergents for low-temperature washing at 30 °C and 40 °C. The results show that while many detergents performed equally well at both temperatures, others showed statistically worse performance at lower temperatures, thus indicating detergent selection might play a role in consumer success with the adoption of energy saving cold-water washing.

2.3 Laundry sustainability consumer factors

Despite the progress made in improving the sustainability of washing machines and detergents, consumer behavior can and do counteract these benefits [42]. Consumers are in near-complete control of the all-important in-use factors: wash temperature, cycle time, load size, and detergent dosing [22]. Many of these consumer-dependent variables have significant environmental consequences [42]. Rationally, consumer behavior should be directed by factual information provided by key stakeholders, but in reality it is difficult to influence consumer behavior, especially when dealing with routine daily tasks [43]. Throne‐Holst, Strandbakken and Stø [44] detailed the barriers to adopting environmental changes and identified cultural norms and economic concerns as being equal to or more important than information availability. In addition, when a task is routine or habitual, consumers typically invest minimal cognitive effort in decision making. While visible stains and strong malodors are clear physical reasons for washing clothes, for many consumers habit is a principal motivator for washing clothes [45].

In addition to its functional benefit, cleanliness is a cultural construct. The decision to wash clothing is motivated by physical needs, habit, emotion, and perceived community censorship making any change in perceived performance emotionally charged [45]. For many consumers, the elimination of malodor and clean, fresh-smelling clothing is an important part of public self-presentation [46]. The absence of malodor and clean scent are significant motivators for washing clothing. In a survey of 2000 US respondents who had purchased and used laundry, cleaning, and air freshening products in the previous six months, Herz et al. [47] found that 76 % of consumers agreed that smell was the signal of a job “well done” when cleaning clothing, which was significantly higher than the signal they obtained from visual inspection. Also, 36.4 % of consumers stated they would rewash clean clothing if it does not smell clean. For others, clean-smelling laundry may elicit feelings of comfort and security [45]. Concerns about clothing scent and public self-presentation are not unfounded, as Kerr, Rosero and Doty [48] found that when posed with a hypothetical person whose clothing was associated with a canonically clean scent, panelists rated the person as more intelligent, attractive, successful, and sociable than someone whose clothing was associated with a scent not consistent with the concept of clean. Thus, following heretofore successful clothes-washing routines provides consumers with peace of mind and anything that raises doubts about cleaning results could make consumers uneasy [49].

2.4 Research questions

This paper investigates the impact of detergent technology on sustainable behavior in the context of a socio-technical system via an interventional study where study participants (henceforth referred to as “subjects”) currently using a detergent marketed as “heavy duty detergent” were given either a similar “heavy duty” detergent (HDD) or an “eco-brand” detergent (EBD) designed to meet Ecolabel criteria. The study tracked the differences in subject acceptance, perceptions, and attitudes towards the two detergents as well as examined their clothes-washing procedures for any habit changes resulting from the change in detergent. Based on the literature discussion, the following questions are posted:

1. Consumer satisfaction

   1. Is there a difference in overall consumer acceptance of the two detergents?

   2. What differences in performance are perceived by the subjects?

2. Do washing behaviors vary between the two test detergent legs? Specifically,

   1. Detergent dosing

   2. Wash temperature

   3. Usage of additional products besides detergent

3. Do consumers consciously attribute any behavior changes to the test detergent?

3.1 Study design

The study was a longitudinal blinded in-home use and diary study comparing two liquid tablet detergents: one test product comprising a detergent designed to comply with EU Ecolabel certification criteria (EBD – eco brand detergent) and a control product, which was a regular high performance laundry detergent optimized for in-use performance (HDD – heavy duty detergent) with 69 French consumers in each leg. The EU Ecolabel requires a fitness for use of those ‘eco-branded’ detergents in requesting that they shall have a satisfactory wash performance at the lowest temperature and dosage recommended by the manufacturer for the water hardness in accordance with ‘EU Ecolabel protocol for testing laundry detergents’ [50]. However, when technically tested against each other, the eco-brand detergent shows lower technical cleaning performance (see Supplementary Material 5).

The study started with a baseline monitoring period with subjects using their own detergent product for 4 weeks to examine established habits (Table 1). At the beginning of the intervention phase, subjects were given either the EBD or HDD in blinded containers (Figure 2) to use and be monitored for 8 weeks (weeks 5–12). After the intervention phase there was a post-test cool down period where subjects returned to their usual product (weeks 13–18). After this cooldown, post-test qualitative interviews were conducted with select subjects. The subjects did not receive information about the importance of washing at lower temperatures from a sustainability point of view.

Figure 2:

Product study blinded packaging and labels. Test product (EBD – coded “FDL” for distribution to subjects) and control product (HDD – coded “MBH” for distribution to subjects), both packed into same pack.

3.2 Subject recruitment

This research was conducted in France, with primary household purchasers and users of laundry detergent (henceforth referred to as “detergent users”) chosen from different parts of the country. The subjects were recruited in February 2022 from a panel owned by a professional agency (La Maison du Test) through the internet. The subjects received monetary incentives for their participation. The following recruitment criteria were applied:

1. All subjects needed to be current users of leading detergent liquid tablet product from a brand marketed as high-performance. This was to ensure study subjects were familiar with the liquid tablet form and a high level of performance.

2. Detergent users who used laundry additives in more than 25 % of loads were eliminated from consideration to prevent additive use from masking detergent performance. Laundry additives encompass anything placed in the washing machine in addition to detergent – i.e. softener, perfume beads, stain removers, pre-treatment products, etc.

3. Subjects were required to have a dynamic washing temperature habit, meaning already using more than one wash temperature. The majority (73 %) of potential subjects in the recruiting panel met this criterion.

3.3 Monitoring/data collection

3.4 Laundry products used during the study

Both products used in the study were characterized in a laboratory test for cleaning performance using the principles from EU Ecolabel Protocol for testing Laundry Detergents [18]. Stain removal was tested on the standard AISE stain set, over a 30 °C cotton short cycle, as per standard methodology [52]. The control (HDD) product cleaned significantly better in 12 of 14 stains versus test (EBD) product. Whiteness was analyzed based on D-65 W Ganz measurement, with the swatches washed in the control (HDD) product being significantly whiter versus the test (EBD) product (+16.7 points after four washes on cotton, +18 points after four washes on cotton/polyester – a delta of five points is consumer noticeable). Freshness intensity was measured on wet & dry fabrics using an ASTM standard quantitative descriptive analysis with trained sensory panelists [53]. The control (HDD) product was fresher smelling versus the test (EBD) product (+7.5 points on smell of wet fabric and +25.6 points on smell of dry fabric). See Supplementary Material 5 for a full technical characterization of the two products.

The test (EBD) products were sourced from the market and the control (HDD) product directly from the manufacturer’s facility. Both products were repacked into identical white tubs for blind unbranded placement (Figure 2), with generic labels including standard dosing instructions, required safety information, and a randomly generated product code. The products were distributed by the local recruitment agency to subjects, including a placement letter to explain the process of the study.

3.5 Measures and statistical analysis

3.5.2 Wash behaviors

For each load, subjects reported in an online diary regarding the conditions of the wash load. This data was analyzed for the base line period and then over the 8 weeks of study product usage to examine potential changes in laundry habits arising from the study products and to look for potential underlying confounding factors.

1. Fill level – Subjects were asked to estimate the fill level of their washing machine on a scale from 1–6 (Figure 3). For analysis, the levels were aggregated into two categories: full (1, 2, 3) and not full (4, 5, 6).

2. Soil load – Subjects were asked to input the soil level of their load. For analysis these were aggregated into either “no/little soil” or “moderately to heavily soiled”.

3. Cycle duration – Subjects were asked to input the duration of their cycle. For analysis, the durations were then aggregated into two groups – under 75 min (considered a shorter cycle) and over 75 min (considered a longer cycle).

4. Use of additives (stain removers, softeners, perfume beads) – Subjects were asked to input all the products used in their loads beside detergent by selecting from a list of potential additives.

5. Wash temperature – Subjects were asked to fill in the temperature of the wash chosen for the load (based on the temperature indicated on the machine dial).

6. Detergent dosing – Subjects recorded the number of laundry pods used for each load.

Figure 3:

Fill level illustration from load diary.

The percent of washes for each parameter per subject were compared between the control (HDD) leg and test (EBD) leg in absolute values through mixed model. Since each individual subject did many washes during the baseline period (4 weeks) and intervention period (8 weeks), all loads and weekly data are auto-correlated, it is appropriate to apply a repeat measure model for behavioral data. Test product, study week, study week order, and two-way interaction of these factors were treated as fixed effects, baseline average as covariate variable, and participant and study week as repeat effects. The primary outcome was least squares mean for each test product. Student’s t was applied as post hoc analysis for products pairwise comparison in different weeks. The Mixed ANOVA model in JMP 16.1 (100 SAS Campus Drive, Cary, NC 27513-2414, USA) was applied.

4.1 Data set

118 subjects completed all 3 phases of the study – 4-week baseline and 8-week intervention, – with 59 in each leg (EBD and HDD). This represented a total of 5300 loads of laundry monitored (1716 during the baseline period with their product of choice and 3584 during the intervention phase). 17 subjects also participated in qualitative interviews, both during and after the end of the study.

4.2 Overall and attribute ratings

The comprehensive surveys at the end of the baseline phase when both legs (EBD and HDD) used their own usual high-performance detergent there was no statistical difference in OAR between legs (Table 2). At the end of the intervention phase the test (EBD) leg showed a statistically significant (p < 0.05) decline in OAR both versus their baseline and versus the control (HDD) leg. Figure 4 compares the distribution of responses with significantly more “very good” and “excellent” responses for the control (HDD) leg. Along with the reduction in OAR the test (EBD) leg is rated statistically significantly lower in both cleaning performance i.e. cleaning, stain removal, whiteness and freshness attributes i.e. freshness, odor removal, and scent (Table 2). The freshness a detergent provides is influenced by many factors such as perfume composition, the amount of perfume, the cleaning performance and other formulation parameters such as the surfactant system and deposit aids. Eco-labelled products often differ for these parameters from detergents without an eco-label and may explain why the smell of clean is noticeable lower.

Figure 4:

Comparison of distribution of overall ratings (OAR) after 8-week intervention by product leg.

4.3 User attitude responses

In addition to reduced overall rating, the subjects using the test (EBD) product also claimed to be less likely to purchase the product and were less likely to trust the product, feel pride in their laundry and confident in the quality of outcome they expect versus the subjects using the control (HDD) product (Table 3).

4.4 Washing behaviors

Six characteristics about individual subject loads were analyzed over the intervention period, comparing them between the control leg and test leg. These differences are reported in Table 4.

In Table 4, two key behaviors showed significant differences (p < 0.05) between the control (HDD) and test (EBD) leg on average over the 8-week intervention period: detergent dosing and wash temperature, with the test (EBD) leg subjects increasing detergent use and higher washing temperature. The other parameters did not show a significant difference, meaning that the different performance of the detergents triggers the consumer to adjust mainly the washing temperature and the amount of detergent dosed.

To better understand these compensating behaviors, average wash temperature and dosing were further analyzed over time, looking at data from every load.

4.5 Dosing of detergent

Examination of the data by individual intervention week shows a trend to increasing dosage for the test (EBD) product over the 8-week period without a plateau. The dosage of the control (HDD) product was much more stable after an initial increase at the beginning of the study (Figure 5). The initial increase on both products can be explained by the test conditions, in which the subjects receive free product to test.

Figure 5:

Mean pods dosed per load over time during the intervention phase.

If we look at the 8 weeks of testing as 2 periods of 4 weeks (Table 5), the average dosing for the control leg remains stable over the 2 periods of the intervention phase (weeks 5–12). For the test leg, the dosing progressively increases, up to a 0.2 tablet per wash significant (p < 0.05) difference between both legs, meaning one more tablet every five washes.

4.6 Washing temperature

The difference in average washing temperature between control leg and test leg also grows over time, from a 0.6 °C in the first 4-week test period to a 1.1 °C higher temperature for the test (EBD) versus control (HDD) leg in the second period (Table 6).

Another way to look at washing temperature is to focus on the percentage of loads that are done in cold i.e. 30 °C or below. The percentage of these loads among all loads were analyzed for both legs on both periods of the intervention phase (Table 7). In the first period of the test (weeks 5–8), both legs do similar levels of cold loads. However, in the second part of the test (weeks 9–12), the control leg increases their proportion of cold loads again while the test leg remains at similar level.

4.7 Interview qualitative results

Interviews were conducted with 18 subjects at either week 11 or week 18 of the study for in-depth qualitative understanding. Some example behaviors included subjects doing more/less cold wash versus their habits, starting to use additives, or rewashing. The interviews enabled linking quantitative behavioral data to the subjects’ insights and explanations behind them. Some quotes from these interviews are included in the discussion section to illustrate key insights.

5.1 Performance differences are noticeable and reduce consumer satisfaction and trust

The first important finding of this study is that the subjects notice the performance difference between HDD and EBD detergents. This is reported via a significant decline for the EBD detergent OAR (Figure 4). This dissatisfaction with EBD performance was seen in ratings for visual cleaning attributes i.e. stain removal and whiteness and was even more pronounced for the “smell of clean” i.e. odor removal and freshness performance attributes (Table 2). The reduced satisfaction led to a significant decrease in purchase intent for the EBD product (Table 3). This is consistent with a study by Shim, Shin and Kwak [55] that found that eco-labeled detergent products which did not deliver equal performance to other detergents resulted in a significant reduction in intent to purchase. Furthermore, test subjects rated the EBD product as less trusted versus the HDD product (Table 3, 46 vs. 73, p < 0.0001). Deficient performance over time may result in loss of trust in labels. Assumptions of having to accept reduced product quality, loss of convenience, or additional expense are a relevant barrier for many consumers to adopt EBD-like products and services [56, 57]. Therefore, poor performance could continue to erode the trust in eco-products as suggested by Shim and Yoshi [55, 57].

5.2 Reduced detergent performance results in consumer compensatory behaviors

The second important finding of this study is that the subjects using EBD products compensate for lower cleaning performance by changing their laundry behavior. The following discusses the two compensatory actions observed: increasing detergent dosing and increasing wash temperature.

6.1 Cold-water washing habit formation and complacency

Given the potential impact of cold-water washing, switching consumers to use more cold-water washes continues to be the holy grail of laundry sustainability [37]. Encouragingly, some geographic regions such as Spain as well as in many other regions outside of Europe already have transitioned to mostly cold-water washes [58]. Research into cultural differences may provide valuable insight into how to overcome the barriers of habit, beliefs, and complacency more broadly.

Even when consumers can see clear differences in performance between detergents, changing washing temperature habit is very slow as also seen in this study and longer test periods should be considered to understand if the habits between EBD and HDD users continue to differentiate further. For many consumers, laundry detergent is a low involvement category, meaning there is low loyalty and frequent brand switching [59]. It may be that the variability in detergent performance from brand to brand does not build consumer confidence in cold-water washing. In this case, it might be useful to research the potential benefit to mandate stringent performance criteria at energy saving cycles also for EBD products, so that consumers are encouraged consistently to build a cold-water wash habit.

Considering the overall lifecycle impact of the laundry socio-technical system, the question needs still to be answered where the absolute acceptable minimum environmental impact is and how it can be achieved. As this intervention study has shown a reduction in performance, due to less and/or lower performing ingredient composition is compensated by consumers by increasing the amount of detergent and using higher wash temperatures, actions that increase the overall environmental impact. This trade-off between detergent performance and compensatory behavior needs to be quantified via LCA to demonstrate lower environmental impacts for the use of heavy duty high performance detergents at lower dose and temperature instead of using lower performing eco-brand products.