Service quality in South African banking builds loyalty mainly through satisfaction and trust, not directly

Introduction

Retail banking in South Africa has moved a long way from the branch queue, though it has not left it behind. Ramsundra and Mason (2024) surveyed 400 digital banking users in the iLembe district of KwaZulu-Natal and found that online service quality, usability and branch service quality were the strongest drivers of digital use. Branch service quality, in other words, still showed up in a study of people who bank on their phones. That small detail says a lot about how messy the question of service quality has become, because customers now judge one bank through several channels at once (Kiliswa and Mugova, 2025).

This paper argues that service quality matters in South African banking, but mostly as a route to customer satisfaction, and that satisfaction and trust carry most of the weight when it comes to loyalty. The older SERVQUAL logic, which treats a bundle of service dimensions as the thing to be measured, needs rewriting for digital delivery, and the South African evidence suggests the rewrite is only half done (Zouari and Abdelhedi, 2021; Gazide, Mokoena and Chauke, 2025). Trust, meanwhile, behaves less like an automatic reward for good service and more like a slow-building background condition (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025). The sections below look at how service quality has been measured and extended, then at what links quality to satisfaction and loyalty, then at trust and reputation, and finally at the limits of the evidence.

From SERVQUAL to digital service quality

The five-dimension SERVQUAL model has been the default way of thinking about service quality in banks for a long time, and it is still the reference point that extended models start from (Zouari and Abdelhedi, 2021; Gazide, Mokoena and Chauke, 2025). The trouble is that it was designed around a human encounter across a counter. Zouari and Abdelhedi (2021) tried to repair that by adding digitalisation and compliance to the original dimensions, and they tested the extended model on 145 customers of Tunisian Islamic banks. Their factor analysis did not return the original structure at all. It produced five different dimensions, which they labelled confidence, compliance, digitalisation, tangibles and human skills, and all of them except tangibles were positively related to satisfaction. The sample was small and came from one country, so this is a pointer rather than a rule, but it does suggest that customers carve up service in ways the original scale never anticipated.

Studies of electronic banking tell a similar story, though the details shift from place to place. Ayinaddis, Taye and Yirsaw (2023) surveyed 385 customers in Ethiopia and found that responsiveness, reliability, security and privacy, speed and convenience each affected satisfaction. System availability, ease of use and service charges did not. Reliability and responsiveness are old SERVQUAL ideas in new clothes, while security and speed are basically additions. What stands out is the finding on ease of use, because Ramsundra and Mason (2024) found the opposite in South Africa, where usability had the largest positive influence of all the drivers they tested. The two studies measure different outcomes, one asking about satisfaction and the other about use, which may explain part of the gap. Even so, nobody should assume that a dimension which matters in one market carries across to another.

The South African online studies report the explanatory power of e-service quality mostly as a set. Gazide, Mokoena and Chauke (2025) surveyed 274 internet banking users drawn from five major banks and reported that the selected e-service quality dimensions together explained 59.1 per cent of the variance in e-satisfaction. Redda (2023) worked with 310 online banking users and likewise found that e-banking quality had a positive effect on satisfaction. These are respectable numbers. They also come from people who already use online banking, and in the five-bank study from convenience sampling, which means they say little about customers who tried the app and went back to the branch, or never tried it.

Where does that leave SERVQUAL? Probably as a starting checklist and not a finished instrument. Its core idea, that quality is the gap between what customers expect and what they get, still looks sound, and nothing in the studies discussed here overturns it (Zouari and Abdelhedi, 2021). What the digital evidence shows is that the dimensions need to be rebuilt for the channel, with security, speed and convenience written in from the start, and that the human parts of the service do not disappear. Zouari and Abdelhedi (2021) still found a factor for human skills in an increasingly digital bank. Ramsundra and Mason (2024) found that branch service quality kept its place among the main drivers of digital use. So a South African bank that measures only its app is seeing part of what its customers experience.

Why satisfaction does most of the work

The clearest South African evidence for the mediating role of satisfaction comes from Van Deventer and Redda (2023). They surveyed 271 Generation Y retail banking customers and built a structural equation model. The path from service quality to loyalty was positive but small and not statistically reliable, with a coefficient of 0.097 and a p-value of 0.175. The path from satisfaction to loyalty, by contrast, came out at 0.793. If those figures hold up elsewhere, and Kiliswa and Mugova (2025) give some reason to think they might, good service on its own does not buy loyalty. It buys loyalty only to the extent that it leaves the customer satisfied, which is a narrower claim than most bank marketing materials would like to make.

Redda (2023) did not find quite the same thing. In that study of 310 online banking users, e-banking quality affected loyalty through satisfaction, but the direct effect of quality on loyalty stayed significant after satisfaction entered the model, so the mediation is described as partial. Put the two studies side by side (Redda, 2023; Van Deventer and Redda, 2023) and there is a real disagreement. One finds no reliable direct link and the other finds one. The samples differ, with Generation Y customers of retail banks in one case and online banking users in the other, and the quality measures differ too. A direct effect may well exist for digital channels, where a failed transfer or a dropped login is hard to forgive, and fade for the general run of branch-based service. That is a guess, though, and neither paper tests it. Gazide, Mokoena and Chauke (2025) at least show that e-service quality explains a lot of digital satisfaction, which fits the guess without proving it.

Two further studies from Durban point towards the indirect route. Kiliswa and Mugova (2025) surveyed 466 bank customers and used partial least squares modelling to show that customer experience affects loyalty indirectly, through service quality and satisfaction. Mugova, Kiliswa and Mashizha (2025) reached a parallel conclusion for multi-channel retail banking, reporting that service quality shapes loyalty through customer satisfaction and that satisfied customers show stronger loyalty across channels. Both studies report 466 responses from bank customers in Durban, so they appear to draw on the same or an overlapping dataset. That is worth keeping in mind, since two papers do not make two independent confirmations. Read as one body of evidence, the Durban work supports the mediation story, and read as two, it risks overstating how settled the question is.

The wider African and emerging-market literature is broadly in line. Ayinaddis, Taye and Yirsaw (2023) reported that satisfaction had a significant effect on loyalty in Ethiopian e-banking, and Gazide, Mokoena and Chauke (2025) found that e-satisfaction explained 74.5 per cent of the variance in reuse intention among South African internet banking users. The consistency is reassuring up to a point. But reuse intention is a statement about the future, not a record of what people actually did, and most of this research measures attitudes at a single moment. A customer who says she will keep using the app and a customer who has stayed with the same bank for fifteen years because switching is a hassle look identical on a questionnaire. Genuine switching data from South African banks, which would settle a lot of this, does not appear in any of the ten sources.

Trust, reputation and what they add

Trust is often bundled with loyalty as if the two were one thing, and the South African evidence suggests they are better kept apart. In the Van Deventer and Redda (2023) model, trust was predicted by satisfaction (0.715) and by commitment (0.257), not simply handed down from service quality. This fits an intuitive picture in which trust builds up over repeated encounters that went well enough, and it echoes the way satisfaction pulled loyalty in the same model (Van Deventer and Redda, 2023). A customer does not trust a bank because the app loaded quickly once. She trusts it after the fifth or sixth time nothing went wrong, and probably after one occasion when something did go wrong and was fixed. The model cannot show that sequence directly, since it is a cross-sectional snapshot, but the pattern is compatible with it.

Rouse, Bátiz-Lazo and Carbó-Valverde (2025) approached trust from the other side, asking whether it drives behaviour. They used ten years of FinScope survey data (a national household survey of financial behaviour) and logistic regression to examine mobile banking adoption in South Africa. Trust in banks had a positive influence on adoption, but its practical impact was modest (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025). Income mattered more, raising the probability of adoption by 9.44 per cent, and adopters were more likely to be aged 30 to 39 and to have advanced education. Black respondents showed higher adoption rates than other ethnic groups. Two things follow, at least as far as one study can show. Trust is not irrelevant, and it is also not the main thing standing between a customer and the app. The finding is also a useful reminder that service quality research, which typically surveys existing users, is looking at the end of a process that income and education shaped long before.

Reputation sits behind both. Van Heerden and Badenhorst (2022) replicated an earlier study of the corporate image of South African banks and again found that behavioural and visual identities both contribute to it. Their eight factors included social interaction, service, quality and reputation, and they concluded that banks should nurture social relationships with customers. They measured image and not trust, so the link is an inference on my part, but it is a reasonable one. Image is what a customer carries into an interaction before anything has been delivered, and it probably colours how a late payment or a slow call-centre queue gets interpreted. A bank with a warm image may be forgiven a bad week, while a bank with a cold one may not, although no study here tests that directly (Van Deventer and Redda, 2023).

Taken together, the trust evidence qualifies the simple chain of quality to satisfaction to loyalty. Satisfaction feeds trust (Van Deventer and Redda, 2023), trust has a modest influence on adopting new channels (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025), and reputation shapes the way service gets read (Van Heerden and Badenhorst, 2022). None of these steps runs on its own. So a bank that wants to understand loyalty has to track at least three things at once, the experience customers actually have, how they feel about it afterwards, and the standing the bank had before they walked in.

What the evidence can and cannot carry

The studies reviewed here share some features that limit how far their conclusions travel. Most are cross-sectional surveys. Several use convenience sampling, including the internet banking study of five major banks (Gazide, Mokoena and Chauke, 2025) and the Tunisian study of 145 customers (Zouari and Abdelhedi, 2021). Samples cluster in particular places, with Durban, iLembe and a Generation Y cohort standing in for a country of very different customers (Van Deventer and Redda, 2023; Ramsundra and Mason, 2024). Nearly everything is self-reported. The single longitudinal design in the set is the FinScope analysis, and that one concerns adoption and trust rather than service quality (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025). So the claim that quality works through satisfaction is well supported as a pattern in survey responses and much less well supported as a description of how customers behave over time.

Inclusion is the other gap, and it is particularly South African. Customers who answer a questionnaire about online banking quality have already crossed the digital threshold, and the people who have not are invisible in the data. The FinScope results suggest that the threshold is not evenly distributed, because adoption rose with income and education (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025). If a bank improves its app and its satisfaction scores rise, that may say as much about who is answering as about who is being served. Branch service quality may be doing quite different work for customers who rely on it, and the fact that it still featured among the leading drivers of digital use in KwaZulu-Natal makes it hard to write off (Ramsundra and Mason, 2024).

A third problem is conceptual. Loyalty is measured in these papers as a mix of intention, attitude and commitment, and the studies do not always say which. Van Deventer and Redda (2023) treat commitment as a separate construct that feeds trust. Gazide, Mokoena and Chauke (2025) measure reuse intention, which is different again from either attitude or behaviour. Until the measures are lined up, comparing coefficients across studies is shakier than the tidy tables suggest. A fair reading is that the direction of the relationships is fairly consistent while their size is not.

What would help most is not another survey of app users. It would be a panel design that follows the same South African customers across channels over a couple of years, records actual switching and product holding, and includes people who use branches, agents and phones in different mixes, which the ten years of FinScope data only partly approximate (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025). Anything short of that will keep producing models that fit well, like the one reported by Van Deventer and Redda (2023), while leaving open whether the explanation holds for customers who are not in the sample.

Conclusion

Service quality does matter in South African banking, though not in the way a plain reading of SERVQUAL would suggest. The evidence points to satisfaction as the working link between what banks deliver and whether customers stay (Van Deventer and Redda, 2023; Mugova, Kiliswa and Mashizha, 2025). A direct effect of quality on loyalty appears in some digital samples and not in others, and the disagreement between studies has not been resolved (Redda, 2023). The original dimensions need to be rebuilt around security, speed, convenience and digital reliability, while the human side of service stays in the measure (Ayinaddis, Taye and Yirsaw, 2023; Zouari and Abdelhedi, 2021).

Trust looks like a slower variable, built from satisfying encounters and from the image a bank already holds, with a real but modest influence on whether customers take up new channels (Rouse, Bátiz-Lazo and Carbó-Valverde, 2025; Van Heerden and Badenhorst, 2022). The practical lesson for a bank manager is to treat satisfaction as the outcome to watch and to measure quality channel by channel (Kiliswa and Mugova, 2025). The lesson for researchers is that the current evidence base, built on convenience samples and stated intentions, is thinner than the volume of publications suggests, and that customers who do not use digital channels remain the biggest blind spot.

Reference list

Ayinaddis, S.G., Taye, B.A. and Yirsaw, B.G. (2023) Examining the effect of electronic banking service quality on customer satisfaction and loyalty: an implication for technological innovation. Journal of Innovation and Entrepreneurship, 12, article 22. doi:https://doi.org/10.1186/s13731-023-00287-y

Gazide, L., Mokoena, A. and Chauke, X.D. (2025) Factors influencing customer satisfaction and reuse intention of internet banking in South Africa: a study on the selected e-service quality dimensions. EUREKA: Social and Humanities, (5), pp. 22-38. doi:https://doi.org/10.21303/2504-5571.2025.003835

Kiliswa, N.G. and Mugova, S. (2025) Customer experience and customer loyalty in retail multichannel banking: a mediation analysis. European Journal of Management Issues, 33(1), pp. 14-22. doi:https://doi.org/10.15421/192502

Mugova, S., Kiliswa, N.G. and Mashizha, M. (2025) The mediating role of customer satisfaction in the relationship between service quality and customer loyalty in multi-channel retail banking. Cogent Business & Management, 12(1), article 2514937. doi:https://doi.org/10.1080/23311975.2025.2514937

Ramsundra, A. and Mason, R.B. (2024) Factors influencing the use of digital personal banking in South Africa. International Journal of Electronic Commerce Studies, 15(3), pp. 29-58. doi:https://doi.org/10.7903/ijecs.2340

Redda, E.H. (2023) E-banking quality and customer loyalty: the mediating role of customer satisfaction. Banks and Bank Systems, 18(2), pp. 177-188. doi:https://doi.org/10.21511/bbs.18(2).2023.15

Rouse, M., Bátiz-Lazo, B. and Carbó-Valverde, S. (2025) Trust and socio-demographic influences on mobile banking adoption in South Africa: a longitudinal study. International Journal of Bank Marketing, 43(10), pp. 2112-2135. doi:https://doi.org/10.1108/IJBM-08-2024-0458

Van Deventer, M. and Redda, E.H. (2023) Customer loyalty and trust in South African retail banking. Innovative Marketing, 19(2), pp. 211-222. doi:https://doi.org/10.21511/im.19(2).2023.17

Van Heerden, C. and Badenhorst, J.J. (2022) Revisiting the corporate image of South African banking institutions. Communicare: Journal for Communication Studies in Africa, 23(1), pp. 15-32. doi:https://doi.org/10.36615/jcsa.v23i1.1780

Zouari, G. and Abdelhedi, M. (2021) Customer satisfaction in the digital era: evidence from Islamic banking. Journal of Innovation and Entrepreneurship, 10, article 9. doi:https://doi.org/10.1186/s13731-021-00151-x