The purpose of this research is to determine whether loyalty programmes can encourage prosocial behavioural loyalty. Following the launch of the Lifeblood Gifts loyalty programme, the authors evaluated its impact on donation frequency, retention and reactivation of voluntary blood donors in Australia.
A quasi-experimental pre-test–post-test research design was used to compare donation counts across three six-month periods: 12 months before, 6 months before and during the trial of Lifeblood Gifts (28/11/2022–28/05/2023). A matched control group (n = 69,337) who did not participate was compared to a group of donors (n = 69,218) who registered for Lifeblood Gifts. Analysis of covariance and t-tests were used for comparisons.
For the intervention group, total donations (33%) and donation frequency (30%) increased during the trial, whereas control group donations and frequencies remained stable. This represented an additional 37,461 donations, mostly plasma. Retention and reactivation rates improved, with 66.2% of first-time donors (control = 24.5%) returning to make a second donation, and 81.3% of lapsed donors (control =26.5%) making at least one donation during the trial period.
Self-selection bias, a short-term focus and potential cultural variability limit the generalisability of the findings.
This study provides novel evidence on the impact of loyalty programmes to encourage prosocial behavioural loyalty and represents an important step forward in understanding how loyalty programmes can contribute to societal health and well-being goals, particularly a more stable supply of blood and blood products.
Loyalty programmes, defined as any “institutionalized incentive system that attempts to enhance consumers’ consumption behavior over time” (Kim et al., 2021, p.73), are an important marketing tool to promote retention and build relationships with customers (Shelper et al., 2019). This definition underscores that, although customer loyalty often includes both attitudinal (brand liking and/or a desire to recommend) and behavioural (changes to actual or intended behaviour) dimensions (Dick and Basu, 1994; Belli et al., 2021), loyalty programmes are primarily designed to influence behavioural loyalty (Melnyk and Bijmolt, 2015). Belli and colleagues in a meta-analysis, found strong evidence that loyalty programmes enhance behavioural loyalty, but shifting attitudinal loyalty is reportedly more challenging (Belli et al., 2021).
In 2024, about 90% of Australian consumers were members of at least one loyalty programme, averaging 4.3 programmes (Posner, 2024). Similar high usage rates are seen in the USA (Fruend, 2017), Europe (Bombaij and Dekimpe, 2020) and Asia-Pacific (Mastercard, 2018). However, less than 50% of members are actively engaged in their loyalty programmes (Fruend, 2017; Posner, 2024), reducing overall effectiveness. Theoretical underpinnings of loyalty programme research (see Chen et al., 2021) have focused on status (e.g. signalling, special treatment and social comparison), inertia/habit formation (e.g. behavioural learning theory and reinforcement) and relational (e.g. social exchange theory, emotional attachment and value) mechanisms driving loyalty programme engagement and effectiveness (Henderson et al., 2011; Kim et al., 2021). Furthermore, the effectiveness of loyalty programmes differs systematically depending on programme design and industry characteristics (Belli et al., 2021).
Research has mostly focused on demonstrating the profitable impact of corporate loyalty programmes on customer retention (Fook and Dastane, 2021) and purchase frequency (Minnema et al., 2017; Son et al., 2020), in retail, hospitality, financial and entertainment sectors (Chen et al., 2021; Mastercard, 2018). Emerging research highlights the positive impact of loyalty programmes on individual health and well-being, such as increased physical activity (Plangger et al., 2022) and responsible gambling tool use (Hollingshead and Wohl, 2024). In Canada, the Carrot Rewards app, which rewards healthy behaviours with retail points, has improved vaccine uptake (Dale et al., 2019), physical activity (Mitchell et al., 2018) and mobile health app engagement (Brower et al., 2020). Partnering with corporate loyalty programmes can also boost health-related behaviours; for instance, airline points to increase engagement with an online heart health programme (Liu et al., 2014). However, there is limited research on whether loyalty programmes can be used to encourage prosocial behaviours that are other-orientated (i.e. improving the well-being of others, society and the environment) rather than self-orientated (i.e. improving personal well-being) (Cavanaugh et al., 2015; Mulcahy et al., 2021).
Some organisations are embedding corporate social responsibility initiatives within loyalty programmes by encouraging prosocial choices (e.g. extra points for sustainable purchases and repeat use of towels) or offering prosocial rewards (e.g. option to donate rewards to charity) (Flacandji et al., 2023). Some research suggests that rewarding prosocial choices can improve attitudinal loyalty, but prosocial rewards do not impact behavioural loyalty in self-orientated contexts. For instance, adding green behaviour rewards can buffer the negative effect of preferential treatment on bystanders’ service satisfaction (Liu and Mattila, 2016), but including a charity reward option does not impact joining a corporate loyalty programme or purchase behaviour, though it could have positive branding implications (Eason et al., 2015). Nonetheless, the social impact of loyalty programmes (offering self-orientated rewards) to encourage behavioural loyalty (repeat and frequent engagement) of other-orientated prosocial behaviour remains unclear. Therefore, this research will address the following research question: can loyalty programmes encourage prosocial behavioural loyalty?
Loyalty programmes are widely used in blood donation (e.g. in the USA and Europe; Graf et al., 2024; Koch et al., 2024), as a prosocial context, but lack empirical evaluation. In Australia, Australian Red Cross Lifeblood (Lifeblood) manages all blood and apheresis collections nationally within an exclusively voluntary non-remunerated system. Only about 75% of donors return within 12 months, with lower retention rates for new (58%) and reactivated donors (66%). Active donors give an average of 4.2 donations per year, with 50% donating fewer than three times annually (Thorpe et al., 2020), despite being able to donate whole blood every 12 weeks and plasma/platelets every 2 weeks. Encouraging retention and regular frequent donations are important objectives for maintaining a stable blood supply.
Consistent with other voluntary blood donation contexts (Zeller et al., 2020), Lifeblood’s national donor recognition policy included blood type key rings, milestone badges, certificates and branded items given out intermittently. Loyalty programmes could encourage blood donation by mitigating high perceived costs (e.g. time, discomfort and anxiety; Lee et al., 1999) and low personal benefit. Kim et al. (2021) found loyalty programmes to be more effective in high-cost, low-frequency industries. However, evidence on incentives for blood donation retention and frequency is mixed (Graf et al., 2023). Results are often dependent on the incentive strategy used and the target segment, with incentives generally more successful for new or infrequent donors (Chell et al., 2018). Loyalty programmes are a distinct systematic incentive strategy that is more acceptable to donors than deal promotions (e.g. vouchers) or thank-you tokens (e.g. milestone awards and gifts) alone (Chell et al., 2022). Australian blood donors are generally positive or neutral towards loyalty programmes, with 56.5%–59.1% agreeing local or national loyalty programmes would encourage others to donate, and 51.8%–56.7% viewing Lifeblood more positively if a loyalty programme was introduced (Van Dyke, et al., 2020). However, in the pursuit of new ways to stabilise and increase donation rates, it remains unclear whether donors would donate more often to obtain Lifeblood-branded items through a loyalty programme.
The purpose of this research was to understand whether loyalty programmes can encourage prosocial behavioural loyalty. Theoretically, loyalty programmes are designed to influence behavioural loyalty through status, reinforcement and relational-based mechanisms (Chen et al., 2021), and material, social and psychological rewards can motivate prosocial behaviour (Erlandsson and Dickert, 2024). Therefore, we hypothesise that a loyalty programme will improve:
donation frequency among active blood donors
retention of early career blood donors
reactivation of lapsed blood donors
In doing so, we directly respond to calls for more research on loyalty programmes in different contexts (Chen et al., 2021) and across different consumer segments (Belli et al., 2021). This research provides the first empirical evidence of the social impact of loyalty programmes to encourage prosocial behavioural loyalty within the context of voluntary blood donation – a globally critical behaviour that directly contributes to the Good Health and Well-Being Sustainable Development Goal (SDG3). Furthermore, as Lifeblood operates as the only blood collection organisation in Australia, this provides a unique context to assess the impact of loyalty programmes on prosocial behaviour without needing to consider competitor offerings (Chen et al., 2021; Liu and Yang, 2009).
Method
Overview of the Lifeblood Gifts loyalty programme
A cross-functional team at Lifeblood used a design-thinking methodology to develop and assess a loyalty programme for blood donation (see an outline of activities in the supplementary file). Lifeblood Gifts was, at the time of the trial, a national loyalty programme that allowed any blood donor to opt-in and systematically access branded items (see Figure 1) based on the type (whole blood and apheresis) and number of attempted donations (successful and unsuccessful). Donors could opt-in to the programme via an email link or QR code displayed in donor centres. The programme comprised a three-level whole blood donation reward structure and a four-level apheresis donation (including plasma and platelets) reward structure. When a donation goal was reached, donors either received the item in centre from staff or received the item by mail after redeeming online (see Figure 2). Lifeblood Gifts continues as a key operational activity (see the current programme: Link to Unbox the new Lifeblood GiftsLink to the website of lifeblood).
The image showcases a variety of promotional items supporting blood donation. Centre stage is a white T-shirt with the phrase "Blood donor for life" printed in red. To the left, two tote bags are present: one orange bag with dots and the text "I'm someone's type," and a dark blue bag with red dots. Below the bags, there is a white mug labelled "Coffee is life," with additional text stating "Wait, no, blood is life. Coffee is a close second." Above, a red rectangular sign reads "Get your blood pumping." On the right side, there are two pairs of socks, one in white and the other in navy, both displaying the phrase "Blood donor for life." Lastly, a red lanyard with text promotes blood donation as well.Lifeblood Gifts branded items
Note(s): Reward pathways: whole blood donation, 1 = lanyard, 2 = socks, 3 = tote bag; apheresis donations, 2 = gym towel, 4 = socks, 6 = tote bag or mug, 8 = t-shirt
Source: Used with permission from Lifeblood
The image showcases a variety of promotional items supporting blood donation. Centre stage is a white T-shirt with the phrase "Blood donor for life" printed in red. To the left, two tote bags are present: one orange bag with dots and the text "I'm someone's type," and a dark blue bag with red dots. Below the bags, there is a white mug labelled "Coffee is life," with additional text stating "Wait, no, blood is life. Coffee is a close second." Above, a red rectangular sign reads "Get your blood pumping." On the right side, there are two pairs of socks, one in white and the other in navy, both displaying the phrase "Blood donor for life." Lastly, a red lanyard with text promotes blood donation as well.Lifeblood Gifts branded items
Note(s): Reward pathways: whole blood donation, 1 = lanyard, 2 = socks, 3 = tote bag; apheresis donations, 2 = gym towel, 4 = socks, 6 = tote bag or mug, 8 = t-shirt
Source: Used with permission from Lifeblood
The infographic outlines a donor recruitment process for Lifeblood Gifts. The process begins with an email invitation encouraging donors to opt-in. Following this, a Q R code is presented, which donors can scan to opt-in easily. After opting in, donors book and attend an appointment for whole blood or apheresis donation. They may either have a successful or unsuccessful donation. Finally, donors who reach a specific donation goal can redeem gifts, either in the centre or online, with items sent by post. The process emphasizes that these steps may repeat during the trial.Lifeblood Gifts member journey
Source: Used with permission from Lifeblood
The infographic outlines a donor recruitment process for Lifeblood Gifts. The process begins with an email invitation encouraging donors to opt-in. Following this, a Q R code is presented, which donors can scan to opt-in easily. After opting in, donors book and attend an appointment for whole blood or apheresis donation. They may either have a successful or unsuccessful donation. Finally, donors who reach a specific donation goal can redeem gifts, either in the centre or online, with items sent by post. The process emphasizes that these steps may repeat during the trial.Lifeblood Gifts member journey
Source: Used with permission from Lifeblood
Data collection approach
Lifeblood Gifts was launched on 28 November 2022 and was evaluated over a six-month period (up to and including 28 May 2023). During that time, 84,801 individuals registered. A quasi-experimental pre-test–post-test research design was used to test the effectiveness of Lifeblood Gifts by comparing donation activity between those who opted in (i.e. intervention group n = 69,218) and a random selection of donors who did not opt in (i.e. control group n = 69,337) across three time periods:
the same six-month period one year earlier (28/11/2021 to 28/05/2022 = Time 1);
the six-month period before the trial (29/05/2022 to 27/11/2022 = Time 2); and
the six-month trial evaluation period (28/11/2022 to 28/05/2023 = Trial).
As such, within-subject (Time 1 vs Time 2 vs Trial) and between-subject (Intervention vs Control) comparisons were performed. It was quasi-experimental as donors self-selected into the intervention group, while the control group was a random selection of donors who did not. This loyalty programme and evaluation methodology was approved by the Lifeblood Human Research Ethics Committee (2021#38-LNR) before the programme launch.
Sampling
To determine the impact of Lifeblood Gifts and ensure a matched control sample, 11 donor cohorts were defined based on donation history before the trial (see Table 1). Despite using comprehensive inclusion and exclusion criteria to make the groups demographically and behaviourally similar, motivational differences likely exist between the intervention and control groups (Chell et al., 2025). Macroenvironmental factors, such as seasonal donation behaviour and the COVID-19 pandemic, likely affected both groups similarly, however, Lifeblood’s marketing activity may have motivated groups differently.
Donor cohorts
| Donor cohort | Total donations | Last donation | Donor panel | Donation frequency^ | Intervention (n)* | Control (n)* | Hypothesis |
|---|---|---|---|---|---|---|---|
| First-time donors | 0 | NA | Any | NA | 3,666 | 3,365 | H2 |
| Early career donors | 1–4 | ≤12 months | WB only | ≥1 | 9,457 | 9,491 | H1 and H2 |
| Mixed/apheresis | ≥1 | 5,939 | 5,971 | H1 and H2 | |||
| Low-frequency donors | ≥5 | ≤12 months | WB only | 1 | 2,554 | 2,581 | H1 |
| Mixed/apheresis | 1–2 | 8,217 | 8,293 | H1 | |||
| Mid-frequency donors | ≥5 | ≤12 months | WB only | 2 | 3,571 | 3,598 | H1 |
| Mixed/apheresis | 3–5 | 13,335 | 13,410 | H1 | |||
| High-frequency donors | ≥5 | ≤12 months | WB only | 3 | 4,031 | 4,042 | H1 |
| Mixed/apheresis | 6–10 | 10,619 | 10,719 | H1 | |||
| Lapsed donors | Any | >12 months | WB only | NA | 3,342 | 3,355 | H3 |
| Mixed/apheresis | NA | 4,487 | 4,512 | H3 | |||
| TOTAL | 69,218 | 69,337 | |||||
Donor cohorts are defined by donation history before the trial launch. WB = whole blood. ^Based on the 12 months before the trial, *Sample size after exclusion criteria applied. Differences in sample size between the “Intervention Group” and “Control Group” is due to exclusion criteria being applied in multiple stages
The intervention group included all who registered for Lifeblood Gifts (n = 84,801), excluding those who opted out by 28 May 2023, were over 75 years old, were prospective donors (i.e. have never donated) or therapeutic donors (i.e. individuals with a medical condition that motivates and/or allows more frequent whole blood donations) and/or donors who were permanently or temporarily deferred (i.e. prevented from donating for safety) that reduced their capacity to donate during the trial compared to pre-trial periods (see Table 2). Very-high-frequency donors were also excluded, as Lifeblood Gifts would have little impact on their already frequent donation behaviour. Henderson et al. (2011) argued that loyalty programmes could disrupt habitual behaviours by making them more salient, potentially having an opposing effect on very frequent donors. A total of 69,218 participants were included in the intervention group analyses.
Exclusion criteria for intervention and control groups
| Excluded if | Intervention group | Control group |
|---|---|---|
| Opted in to Lifeblood Gifts on or before 28 May 2023 | ✓ | |
| Opted out of Lifeblood Gifts on or before 28 May 2023 | ✓ | ✓ |
| Lifeblood staff | ✓ | ✓ |
| [Age] >75 years old at trial launch (maximum age for first-time donors is 75) | ✓ | ✓ |
| [Prospect donor] 0 successful collections recorded by 28 May 2023 | ✓ | ✓ |
| [Therapeutic donor] donor has had at least one therapeutic marked collection or gave >3 successful whole blood collections during the trial evaluation period | ✓ | ✓ |
| [Very-high-frequency donor] donated ≥4 whole blood donations and/or ≥11 apheresis (plasma or platelet) donations in the 12 months before trial lunch | ✓ | ✓ |
| [Permanently deferred] donor had a permanent deferral start before 27 February 2023 (i.e. less than 3 months of eligible loyalty programme participation) | ✓ | ✓ |
| [Temporarily deferred] donor had one or more temporary deferrals that stopped them from donating all product types for the duration of the trial | ✓ | ✓ |
| [vCJD deferred] donor had a vCJD deferral and did not attend a donation appointment before 28 May 2023 (i.e. this deferral is specific to people who lived in the UK during the mad cow disease outbreak [1980 to 1996], and who may have returned to the UK during the trial) | ✓ | |
| [DAE] donor experienced an adverse event at their last donation >2 years before trial launch | ✓ | |
| [Previously deferred and not returned] donor had a temporary deferral applied within the 12 months before the trial and did not attend a donation appointment during the trial period | ✓ | |
| [No appointments] donors did not have at least one appointment during the trial period (low-, mid-, high- and very-high-frequency donors only) | ✓ |
A stratified random sampling approach identified a matched control group of donors who did not opt-in to Lifeblood Gifts (n = 69,337), based on the size of each donor cohort in the intervention group. Additional exclusion criteria ensured the control group had an equal opportunity to donate during the trial and similar engagement with Lifeblood (see Table 2). The lapsed donor control group was randomly selected from donors (≤5 years lapsed) who opened the most recent Lifeblood e-newsletter, reflecting “engaged” lapsed donors rather than “disengaged” ones.
Analysis
Routinely collected data from Lifeblood donor records were extracted for three time periods (Time 1, Time 2 and Trial) and compared to evaluate trial success. The primary dependent variable was the number of successful collections to evaluate the impact of Lifeblood Gifts on donation frequency (H1), early career donor retention (H2) and lapsed donor reactivation (H3). Analysis of covariance, controlling for age, gender, donation history, years since first donation and temporary deferral status before and during the trial, was performed to compare six-month donation frequency between the intervention and control groups across each time period. The t-tests were performed to compare sample characteristics between the intervention and control groups. However, it is important to note that large sample sizes can lead to inflated significance levels; therefore, the practical significance of observed differences should be considered alongside the statistical significance (Hair et al., 2010).
Results
Sample characteristics
A total of n = 138,555 blood donors were included in the analysis (intervention = 69,218 and control = 69,337). Statistically significant differences (p < 0.001) exist between the groups, though not all are practically meaningful (see Table 3). For instance, control group donors are generally older and have been donating slightly longer. More meaningful differences are seen in gender and temporary deferral rates. The control group has a higher proportion of male donors and few temporary deferrals in the 12 months before and during the trial compared to the intervention group. Research shows that temporarily deferred donors are less likely to return (Davison et al., 2020). In addition, male donors are less likely to experience adverse events (e.g. loss of consciousness) and have a lower risk of deferral due to health reasons or unsuitable veins, increasing their likelihood of returning (Carver et al., 2018). These differences suggest the control group would be more likely to donate during the trial period.
Sample characteristics
| Donor cohort | Age in years | Gender | Years since first donation | # donations before trial | Had a temporary deferral (12 months before trial) | Had a temporary deferral (during trial) |
|---|---|---|---|---|---|---|
| Mean (SD) | % | Mean (SD) | Mean (SD) | % | % | |
| First-time donors | ||||||
| INT | 38.0 (14.5) | M = 39.6% F = 60.4% | N/A | N/A | 4.5% | 26.5% |
| CON | 39.0 (14.6) | M = 52.5% F = 47.5% | N/A | N/A | 1.3% | 15.5% |
| t-test | t(1, 7,029) = −2.89 p = 0.004 | t*(1, 6,950) = 6.68 p < 0.001 | t*(1, 5,780) = 8.10 p < 0.001 | t*(1, 6,941) = 11.48 p < 0.001 | ||
| Early career donors | ||||||
| INT | 40.8 (14.2) | M = 38.9% F = 61.1% | 2.7 (4.7) | 2.3 (1.1) | 23.1% | 25.0% |
| CON | 38.3 (14.1) | M = 46.9% F = 53.1% | 2.9 (4.6) | 2.2 (1.1) | 19.0% | 11.5% |
| t-test | t*(1, 30,849) = 15.38 p < 0.001 | t*(1, 30,845) = 14.12 p < 0.001 | t(1, 30,791) = −5.46 p < 0.001 | t*(1, 30,856) = 9.82 p < 0.001 | t*(1, 30,678) = 8.89 p < 0.001 | t*(1, 28,275) = 31.35 p < 0.001 |
| Active donors^ | ||||||
| INT | 44.5 (14.0) | M = 44.5% F = 55.5% | 11.4 (7.0) | 29.9 (31.8) | 37.8% | 24.6% |
| CON | 46.7 (14.6) | M = 54.4% F = 45.6% | 11.5 (7.0) | 28.7 (29.7) | 36.0% | 22.5% |
| t-test | t*(1, 84,883) = −22.77 p < 0.001 | t*(1, 84,966) = 29.24 p < 0.001 | t*(1, 84,935) = −2.07 p = 0.019 | t*(1, 84,484) = 5.38 p < 0.001 | t*(1, 84,941) = 5.57 p < 0.001 | t*(1, 84,839) = 7.39 p < 0.001 |
| Lapsed donors | ||||||
| INT | 40.5 (12.9) | M = 35.7% F = 64.3% | 10.5 (6.6) | 14.7 (21.5) | 18.6% | 22.0% |
| CON | 42.3 (13.7) | M = 52.1% F = 47.9% | 9.0 (6.1) | 12.0 (17.3) | 10.3% | 6.5% |
| t-test | t*(1, 15,640)* = −8.82 p < 0.001 | t*(1, 15,673)* = 20.91 p < 0.001 | t*(1, 15,609) = 14.38 p < 0.001 | t*(1 ,14,972) = 8.48 p < 0.001 | t*(1, 14,792) = 14.88 p < 0.001 | t*(1, 12,719) = 28.58 p < 0.001 |
^Active donors includes low-, mid- and high-frequency donors. *Lavene’s Test for Equality of Variances was violated; therefore, t-statistics based on equal variances not assumed, SD = standard deviation
Donation frequency (H1)
To determine whether Lifeblood Gifts impacted donation frequency, we focused on active donor cohorts with the potential to improve donation frequency over a six-month period, including early career, low-, mid- and high-frequency donors. Within the control group, donation frequency and total successful collections remained stable across all time points (see Figure 3). The slight decrease in the number of active donors during the trial (compared to Time 2) may reflect a lower rate of retention among the control group. Within the intervention group, a substantial increase in the number of active donors (34%) coincides with a substantial increase (39%) in total successful collections between Time 1 and Time 2, but there is minimal change in donation frequency (5%). However, during the trial, despite the number of active donors only increasing by 1%, total successful collections increased by 33% and donation frequency increased by 30% when compared to Time 2. This represents an additional 37,461 collections during the trial, predominantly driven by more plasma donations. Furthermore, when comparing six-month donation frequency between the intervention and control group, there was no significant difference at Time 1 [F(1, 80,595) = 0.64, p = 0.425] but the groups did significantly differ at Time 2 [F(1, 98,285) = 260.80, p < 0.001] and during the trial [F(1, 95,168) = 5,723.27, p = 0.000]. This pattern of results is consistent across all donor cohorts, where the intervention group donated more frequently during the trial than the control group (see Table 4).
Donation frequency across all donor cohorts
| Donor cohort | Condition | Time 1 mean (SD) | Time 2 mean (SD) | Trial mean (SD) |
|---|---|---|---|---|
| First-time donors | Intervention | 2.6 (1.9) | ||
| Control | 1.4 (0.8) | |||
| ANCOVA^ | F(1, 7,026) = 1,253.26, p < 0.001 | |||
| Early career donors | Intervention | 1.3 (0.5) | 1.6 (0.8) | 2.5 (2.0) |
| Control | 1.2 (0.5) | 1.4 (0.6) | 1.6 (1.1) | |
| ANCOVA^ | F(1, 11,955) = 0.00, p = 0.954 | F(1, 246,256) = 558.84, p < 0.001 | F(1, 19,707) = 1,154.86, p < 0.001 | |
| Low-frequency donors | Intervention | 1.2 (0.4) | 1.2 (0.4) | 2.3 (1.8) |
| Control | 1.2 (0.4) | 1.2 (0.4) | 1.5 (0.9) | |
| ANCOVA^ | F(1, 11,978) = 0.57, p = 0.451 | F(1, 13,786) = 1,405.21, p < 0.001 | F(1, 18,427) = 1,405.21, p < 0.001 | |
| Mid-frequency donors | Intervention | 1.9 (0.9) | 2.1 (1.1) | 2.8 (2.1) |
| Control | 1.9 (0.9) | 1.9 (1.0) | 1.9 (1.3) | |
| ANCOVA^ | F(1, 28,761) = 4.89, p = 0.027 | F(1, 30,959) = 253.22, p < 0.001 | F(1, 29,970) = 2,095.43, p < 0.001 | |
| High-frequency donors | Intervention | 3.2 (1.8) | 3.5 (2.0) | 4.0 (2.6) |
| Control | 3.2 (1.8) | 3.2 (1.8) | 2.8 (2.1) | |
| ANCOVA^ | F(1, 27,877) = 29.20, p < 0.001 | F(1, 28,891) = 97.21, p < 0.001 | F(1, 27,040) = 1,728.40, p < 0.001 | |
| Lapsed donors | Intervention | 2.3 (1.8) | ||
| Control | 1.4 (0.9) | |||
| ANCOVA^ | F(1, 8,445) = 491.84, p < 0.001 |
Time 1 = 12 months before trial (28 November 2021 to 28 May 2022); Time 2 = six months before trial (29 May 2022 to 27 November 2022); average donation frequency is based on total successful collections. SD = standard deviation. ^Analysis of covariance controlling for age, gender, donation history, years since first donation and temporary deferral status before and during the trial
In the top left panel, titled 6-Month Average Frequency, the intervention group increased from 2.2 at Time 1 to 3 at Trial, while the control group remained stable at 2.1. In the top right panel, Total Successful Collections rose from 82,227 to 151,876 in the intervention group, while the control group declined from 89,076 at Time 1 to 90,137 at Trial. The middle left panel shows Whole Blood Collections, where the intervention group rose from 33,404 to 44,497 and the control group declined from 37,747 to 36,031. In the middle right panel, Plasma Collections for the intervention group increased from 47,671 to 104,247, while the control group decreased from 49,759 to 52,190. The bottom left panel on Platelet Collections shows a rise for the intervention group from 1,152 to 3,132, whereas the control group increased only slightly from 1,570 to 1,916. In the bottom right panel, the Number of Donors with at least one Successful Collection increased from 38,057 to 51,261 in the intervention group and from 42,518 to 43,917 in the control group. Across all panels, the intervention group consistently shows upward trends, while the control group remains flat or declines slightly.Total successful collection counts of active donors
Note(s): Time 1 = 12 months before trial (28 November 2021 to 28 May 2022); Time 2 = six months before trial (29 May 22 to 27 November 2022); ^six-month average frequency is based on total successful collections
Source: Authors’ own work
In the top left panel, titled 6-Month Average Frequency, the intervention group increased from 2.2 at Time 1 to 3 at Trial, while the control group remained stable at 2.1. In the top right panel, Total Successful Collections rose from 82,227 to 151,876 in the intervention group, while the control group declined from 89,076 at Time 1 to 90,137 at Trial. The middle left panel shows Whole Blood Collections, where the intervention group rose from 33,404 to 44,497 and the control group declined from 37,747 to 36,031. In the middle right panel, Plasma Collections for the intervention group increased from 47,671 to 104,247, while the control group decreased from 49,759 to 52,190. The bottom left panel on Platelet Collections shows a rise for the intervention group from 1,152 to 3,132, whereas the control group increased only slightly from 1,570 to 1,916. In the bottom right panel, the Number of Donors with at least one Successful Collection increased from 38,057 to 51,261 in the intervention group and from 42,518 to 43,917 in the control group. Across all panels, the intervention group consistently shows upward trends, while the control group remains flat or declines slightly.Total successful collection counts of active donors
Note(s): Time 1 = 12 months before trial (28 November 2021 to 28 May 2022); Time 2 = six months before trial (29 May 22 to 27 November 2022); ^six-month average frequency is based on total successful collections
Source: Authors’ own work
Retention (H2)
Among first-time donors (i.e. donated for the first time during the trial), 66.2% in the intervention group returned for a second donation within six months or less and 14.2% donated five or more times (compared to only 24.5% returning for a second donation in the control group). Furthermore, retention rates of early career donors in the intervention group were almost twice as high (>80%) than the control group (<45%) (see Table 5). Given that the control group has a higher proportion of male donors and lower rates of deferrals, we would have expected this group to have a higher return rate (Carver et al., 2018; Davison et al., 2020).
Retention of first-time and early career donors
| Group | # donations before trial | Total (n) | % of returned donors (≥1 collection) | # donations during trial | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 (%) | 2 (%) | 3 (%) | 4 (%) | 5 (%) | 6 (%) | 7 (%) | 8 (%) | 9 (%) | 10+(%) | ||||
| INT | 0 | 3,666 | 100 | 33.8 | 29.8 | 13.5 | 8.8 | 5.6 | 3.9 | 1.7 | 1.5 | 0.8 | 0.7 |
| 1 | 5,484 | 81.2 | 32.7 | 27.8 | 8.0 | 4.5 | 3.2 | 1.8 | 1.5 | 0.7 | 0.5 | 0.5 | |
| 2 | 5,151 | 84.7 | 33.4 | 26.5 | 9.2 | 5.2 | 3.6 | 2.7 | 1.6 | 1.0 | 0.7 | 0.8 | |
| 3 | 3,984 | 83.7 | 30.9 | 25.2 | 9.3 | 5.7 | 3.6 | 3.4 | 2.0 | 1.6 | 1.0 | 1.1 | |
| 4 | 3,484 | 83.9 | 30.5 | 24.5 | 9.2 | 5.8 | 4.4 | 3.5 | 2.4 | 1.5 | 1.0 | 1.3 | |
| CON | 0 | 3,365 | 100 | 75.5 | 18.0 | 3.8 | 1.3 | 0.8 | 0.3 | 0.1 | 0 | 0.1 | 0 |
| 1 | 6,989 | 34.7 | 22.6 | 9.0 | 1.6 | 0.8 | 0.4 | 0.1 | 0.1 | 0 | 0 | 0 | |
| 2 | 4,810 | 40.3 | 24.7 | 10.9 | 2.5 | 1.0 | 0.7 | 0.4 | 0.1 | 0.1 | 0.1 | 0 | |
| 3 | 3,717 | 42.5 | 26.4 | 11.1 | 2.4 | 1.2 | 0.5 | 0.3 | 0.3 | 0.1 | 0.1 | 0.2 | |
| 4 | 3,278 | 43.7 | 26.6 | 12.1 | 2.3 | 1.2 | 0.7 | 0.4 | 0.1 | 0.2 | 0 | 0.1 | |
INT = intervention, CON = control, Donations refer to successful whole blood and/or apheresis (plasma/platelet) collections
Reactivation (H3)
Lifeblood Gifts successfully attracted, reactivated and retained a substantial number of short-term to long-term lapsed donors. Of the 7,829 lapsed donors who registered for Lifeblood Gifts, 6,366 (81.3%) were reactivated by returning to make at least one successful collection during the trial (compared to 26.5% of the control group). Furthermore, 46.9% of lapsed donors in the intervention group returned to make more than one collection and 9.6% returned to make five or more collections (see Table 6). The intervention group had more long-term lapsed donors (>5 years), whereas the control group had more mid-term lapsed donors (2–5 years). Despite a higher proportion of female donors, deferrals and long-term lapses – all factors linked to lower return rates (Kasraian et al., 2020) – intervention lapsed donors were significantly more likely to return and donate multiple times.
Reactivation of lapsed donors
| Group | # years since last donation (pre-trial) | Total (n) | Total # reactivated donors (%) | # donations during trial | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 (%) | 2 (%) | 3 (%) | 4 (%) | 5 (%) | 6 (%) | 7 (%) | 8 (%) | 9 (%) | 10 (%) | ||||
| INT | All | 7,829 | 81.3 | 34.4 | 22.4 | 9.5 | 5.5 | 3.7 | 2.2 | 1.7 | 1.0 | 0.5 | 0.5 |
| Up to 2 | 4,433 | 77.2 | 36.1 | 20.9 | 8.3 | 4.2 | 3.0 | 1.7 | 1.4 | 0.6 | 0.5 | 0.5 | |
| Up to 3 | 985 | 85.8 | 34.9 | 23.8 | 10.7 | 6.9 | 2.6 | 1.9 | 2.8 | 1.2 | 0.5 | 0.4 | |
| Up to 4 | 565 | 85.5 | 32.6 | 24.6 | 11.2 | 6.4 | 4.2 | 3.2 | 1.2 | 1.1 | 0.7 | 0.4 | |
| Up to 5 | 417 | 88.2 | 31.9 | 22.8 | 12.9 | 7.4 | 7.0 | 2.2 | 2.9 | 0.7 | 0.0 | 0.4 | |
| 6+ | 1,429 | 87.3 | 30.1 | 24.8 | 10.8 | 7.9 | 5.5 | 3.4 | 1.7 | 2.0 | 0.6 | 0.6 | |
| CON | All | 7,867 | 26.5 | 19.4 | 5.1 | 1.2 | 0.3 | 0.2 | 0.1 | 0.1 | 0.0 | 0.0 | 0.0 |
| Up to 2 | 3,301 | 37.6 | 27.3 | 7.6 | 1.6 | 0.4 | 0.2 | 0.2 | 0.1 | 0.0 | 0.1 | 0.0 | |
| Up to 3 | 1,855 | 23.1 | 16.8 | 4.0 | 1.3 | 0.3 | 0.3 | 0.1 | 0.1 | 0.1 | 0.0 | 0.0 | |
| Up to 4 | 1,505 | 15.4 | 11.8 | 2.7 | 0.5 | 0.1 | 0.1 | 0.1 | 0.1 | 0.1 | 0.0 | 0.0 | |
| Up to 5 | 1,189 | 15.4 | 10.8 | 3.0 | 0.7 | 0.5 | 0.3 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | |
| 6+ | 17 | 17.6 | 17.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | |
INT = intervention, CON = Control, Donations refer to successful whole blood and/or apheresis (plasma/platelet) collections
Discussion
The purpose of this research was to understand whether loyalty programmes can encourage prosocial behavioural loyalty, within the context of voluntary blood donation. The findings of this study demonstrate that loyalty programmes can, at least in the short-term, improve the extent that people will continue prosocial activity (retention), increase the rate of prosocial activity (frequency) and re-engage following a long period of inactivity (reactivation). Given the increasing importance to create social impact that improves societal health and well-being (SDG3), this study represents an important step forward in understanding how loyalty programmes, as a marketing tool, can contribute to these goals.
Theoretical contributions
This research evaluates the introduction of a nationwide loyalty programme to encourage voluntary blood donation, providing the first empirical evidence of the prosocial impact of loyalty programmes. In doing so, we make two key theoretical contributions. Firstly, we show that offering personal benefits can effectively motivate prosocial behaviour, extending loyalty programme research to actions benefiting others (rather than the individual). Secondly, we validate claims that loyalty programmes influence behavioural loyalty in the form of frequency, retention and reactivation of customers (Belli et al., 2021), and that such impacts are more notable in high-cost low-frequency contexts (Kim et al., 2021) and among less loyal, infrequent customers (Liu, 2007). Donation frequency was higher among all donor cohorts who had opted in to Lifeblood Gifts compared to the control group, with the greatest effect on early career and low-frequency donors (Chell et al., 2018). Together, these contributions provide fundamental support towards broadening the scope and application of existing loyalty programme research to prosocial contexts (Chen et al., 2021).
Practical implications
Loyalty programmes are effective marketing tools for encouraging prosocial behavioural loyalty, but their design characteristics likely impact their effectiveness (Southcott et al., 2022). This research offers practical implications and recommendations for using loyalty programmes as a systematic incentive strategy (Chell et al., 2022).
Firstly, voluntary blood donors donated more frequently to obtain Lifeblood-branded items at set donation intervals, compared to the previous intermittent schedule. A fixed reinforcement schedule seems to further motivate action by creating an expectation for reward (Chell et al., 2022). However, a points-based system may offer more flexibility in what actions are rewarded, such as social media engagement, donor recruitment or responding during shortages and more autonomy in what rewards are redeemed. Given that positive social norms can enhance the effectiveness of incentives for prosocial behaviour (Graf et al., 2023) and that Australia favours points-based loyalty programmes (Mastercard, 2018), this design may be more effective. Economic efficiencies and social impact could also be achieved through a multi-vendor programme (Bombaij and Dekimpe, 2020) or partnerships with existing corporate loyalty programmes (Liu et al., 2014).
Lifeblood Gifts used status (e.g. branded items signalling donor identity; Lam et al., 2021), reinforcement (e.g. rewards at set intervals; Chell et al., 2022) and relational (e.g. rewards positioned as a “thank you” to show donors they are valued; Veldhuizen, 2010) mechanisms to influence blood donor behavioural loyalty (Chen et al., 2021). Further research is needed to identify the most important theoretical mechanisms for optimising prosocial loyalty programmes.
Limitations and future research
The current study provides strong foundational evidence supporting the use of loyalty programmes to promote prosocial behavioural loyalty. While certain limitations limit the generalisability of the findings, there are also opportunities for future research. Despite using comprehensive criteria to make the intervention and control groups demographically and behaviourally similar, the quasi-experimental design may result in motivational difference due to self-selection bias. Loyalty programmes likely appeal to specific segments, such as blood donors who are more deal prone and consider the benefits to outweigh the financial costs (Chell et al., 2025), often leading to an overestimation of their impact on behavioural loyalty as more loyal customers may choose to join the programme to gain benefits. We attempted to control for this by matching the control group based on previous behavioural patterns. However, even when accounting for self-selection, loyalty programmes still enhance behavioural loyalty (Leenheer et al., 2007).
While this research focused on behavioural loyalty, further research is needed to understand whether loyalty programmes can improve attitudinal loyalty to prosocial behaviour and the incentivising organisation (Belli et al., 2021). Evidence suggests that prosocial rewards can improve attitudinal loyalty within corporate loyalty programmes (Liu and Mattila, 2016), but the impact of self-orientated rewards on prosocial attitudinal loyalty remains unclear. This is particularly important in competitive environments, where blood collection systems and donation rates vary internationally (Gorleer et al., 2020). Decentralised plasma collection systems, like those in the USA and Germany, tend to use more diverse incentive strategies to attract and retain plasma donors (Koch et al., 2024). The effectiveness of loyalty programmes is also likely influenced by existing incentive strategies (Graf et al., 2024; Zeller et al., 2020) where behavioural loyalty is anchored to the status quo (Samuelson and Zeckhauser, 1988) and cultural factors (Beck et al., 2015; Graf et al., 2023). Bombaij and Dekimpe (2020) found that retail loyalty programmes are more effective in countries with cultures that are more individualistic and long-term oriented.
This study shows clear evidence of a short-term effect, but it is uncertain whether increased donation frequency is sustainable long-term (Faulkner et al., 2016). Donation frequency can vary based on donors’ commitments and health (Thorpe et al., 2020). Evaluating the long-term effectiveness of loyalty programmes is crucial (Liu, 2007), especially in the prosocial sector where there remain concerns about motivational crowding out (Newman and Cain, 2014). In addition, researchers must consider the impact of a loyalty programme on the attitudinal and behavioural loyalty of both members and bystander donors who do not want to join (Liu and Mattila, 2016; Steinhoff and Palmatier, 2016). Further research is needed to understand the long-term implications of a bifurcated donor panel, where one group is motivated by extrinsic rewards and the other is not.
Acknowledgment
The authors would like to acknowledge Ayesha Kumri for her pivotal role in leading the design-thinking workshops, which were instrumental in transforming Lifeblood Gifts from a concept to full implementation, and for documenting the workshop activities into a practical resource for others to benefit from. The authors would also like to acknowledge the contributions made by Sue Wilkes, Rodney Treble, Claire White, Augustin Raj, Jo Logan, Jo Lawson, Linda Rendal-Jones, Chad Hamp, Batya Atlas, Sarah Hayward, Andrew Jeffrey and Richard Western to the design and execution of the Lifeblood Gifts trial.
