Personalization Paradox: When Brands Know Too Much

Personalization based on artificial intelligence (AI) is often seen as a way to make digital experiences more relevant. However, the relevance of recommendations does not automatically mean that consumers come to trust the platform providing them more.

A study by Purnama and Saragih (2026) of 200 TikTok Shop users in Indonesia found that perceptions of AI-based personalization were negatively related to online trust and electronic loyalty (e-loyalty). Privacy concerns were also negatively related to both variables.

Conversely, online trust was positively related to e-loyalty. The study also found that online trust acted as a partial mediator in the relationships examined.

The research model explained 36.1% of the variance in online trust and 46.1% of the variance in e-loyalty. These figures indicate the model’s explanatory power for the two variables; they are not the percentage of consumers who lost trust or loyalty.

These findings offer an important perspective for brands: personalization and trust do not always move in the same direction. The ability of technology to recognize consumer preferences is not enough to ensure that the resulting experience will be received positively.

But does this mean that AI personalization actually harms consumer trust?

Not necessarily. Other research shows different results.

A study by Rasyiddin et al. (2026) of 248 e-commerce users in Indonesia, collected in February–March 2025 from Jakarta, Bandung, Surabaya, and Yogyakarta, found that perceived personalization was positively related to trust in the platform.

This relationship is reflected in a coefficient of β = 0.442 with p < 0.001. The same study also found an indirect effect of personalization on purchase intention through platform trust, with a coefficient of β = 0.125 and p < 0.001.

These findings suggest that personalization can be associated with higher trust and purchase intention. In other words, personalization does not always provoke a negative response. In the context of that study, personalization was associated with experiences that can support consumer trust in the platform.

Similar results can be seen in the study by Song, Lim, and Oh (2021). Through two experiments with a total of 416 participants, they found that perceived usefulness of personalization was positively related to the intention to use e-commerce apps. However, this relationship was also influenced by privacy concerns and consumers’ willingness to share personal information.

These three studies do not necessarily contradict one another. The context, the variables measured, and the research design need to be taken into account before comparing their results. Nor are these findings sufficient to establish a single cause for why consumer responses to personalization can differ.

Even so, there is one common thread relevant to brands: the value of personalization is determined not only by the technology’s ability to recognize consumers, but also by how consumers evaluate the resulting experience.

Recommendations that feel useful can help consumers find suitable products. However, the more personal an experience becomes, the more important it is for brands to pay attention to how the use of that data is perceived.

If personalization can both help and raise concerns, the next question is what keeps consumers trusting a brand.

PwC Indonesia’s Voice of the Consumer Survey 2026 offers a picture of the importance of trust and transparency in consumers’ relationships with brands. In the survey, 88% of respondents ranked trust as equally important as price and quality, 81% said they were willing to switch if they found better transparency, and 63% expected strong data protection and security.

These figures need to be read according to the survey’s context, namely products, health, and wellness. These data are not a direct measurement of consumer acceptance of AI personalization. Nevertheless, the findings provide context on the importance of trust, transparency, and data protection in consumer decisions.

For brands that rely on personalization, the relevance of recommendations should therefore be viewed together with consumers’ experience of using the service. Consumers may appreciate recommendations that match their needs, but they also want to feel that their personal data is being managed well.

This means that a personalized experience cannot merely feel right. Brands also need to consider whether the way that experience is shaped can be understood and accepted by consumers.

Data protection is an important part of this discussion. In Indonesia, Law Number 27 of 2022 on Personal Data Protection regulates the processing of personal data and the rights of data subjects.

The law was promulgated on 17 October 2022 and set a two-year transition period through Article 74, which ended in October 2024.

For brands that use data for personalization, regulatory compliance is an important aspect of managing consumer data. However, legal compliance and consumer acceptance are two different things.

A practice may meet the applicable provisions, yet not necessarily be perceived by consumers as a comfortable or transparent experience. Conversely, recommendations that feel relevant do not automatically answer consumers’ questions about how their data is used.

For this reason, data protection should not be viewed merely as a compliance obligation. In a personalization strategy, data protection is also tied to how a brand builds and maintains trust.

There is no single level of personalization that can be considered ideal for all consumers and all situations. A recommendation that helps one person may feel too specific to another. Consumer responses can also differ depending on the context of use, the type of data used, and their experience with the platform.

Therefore, brands need to view personalization not only as a matter of accuracy, but also as a matter of consumer acceptance.

In this context, the trust threshold can be used as a strategic framework for understanding the limit at which personalization is still considered beneficial without eroding trust. The term is an interpretive framework to aid decision-making, not a threshold value established by the studies above.

This framework steers brands toward more relevant questions than simply whether a recommendation succeeds in attracting consumer attention: do consumers feel helped, do they understand the experience being provided, and is the way the brand uses data still acceptable?

These questions matter because consumers’ responses to recommendations do not always reflect their overall experience. Interactions or purchases may show that a recommendation succeeded in driving action, but they are not enough to conclude that consumers are also comfortable with the way the personalization is carried out.

For brands, the challenge is to ensure that improvements in personalization performance do not proceed without regard for consumer acceptance. Evaluation can be carried out through three complementary layers.

1. Measure recommendation performance.

Track indicators such as clicks, conversions, and interactions to understand whether personalization helps consumers find relevant products or content. These indicators are important for assessing the effectiveness of the strategy, but they are not sufficient to describe consumer trust as a whole.

2. Monitor consumer acceptance.

Pay attention to indicators such as decisions to turn off personalization, privacy-related complaints, and the results of consumer trust measurements. These indicators can help brands understand how the personalization experience is received, although each needs to be interpreted in context and does not stand alone as proof that consumers have lost trust.

3. Read performance and acceptance together.

When recommendations generate good clicks or conversions, but at the same time privacy complaints rise or trust scores fall, brands need to examine that experience further. Conversely, personalization that is well received but has not yet delivered the expected performance also needs to be evaluated in terms of its relevance and usefulness.

By reading both sides together, brands can avoid concluding that personalization that is commercially effective is necessarily effective in building long-term relationships with consumers.

Ultimately, the challenge of personalization is not merely making AI smarter at recognizing consumers. The challenge is to ensure that this capability produces an experience that genuinely helps, without neglecting privacy and trust.

The more a brand knows its consumers, the greater the opportunity to deliver relevance. However, relevance is only valuable when consumers also accept the way the brand delivers it.

Sources

Purnama, D., & Saragih, F. M. (2026). AI-powered personalization and privacy concern as drivers of e-loyalty: The mediating role of online trust in TikTok Shop. Indonesian Journal of Digital Business, 6(1), 81–93. https://ejournal-tasikmalaya.upi.edu/ijdb/article/view/153

Rasyiddin, A., Rahayu, A., Dirgantari, P. D., Hayati, R., Azis, N., Metris, D., & Romadhon, D. N. A. (2026). AI marketing capabilities and purchase intention in Indonesian e-commerce: Mediating effects of platform trust and AI transparency, and a strategic marketing and digital human resource capability perspective. F1000Research, 15, 1554. https://doi.org/10.12688/f1000research.188717.1

Rasyid, A., Rahayu, A., Dirgantari, P. D., Hayati, R., Azis, N., Metris, D., & Romadhon, D. N. A. (2026). AI personalization, platform trust, and purchase intention survey dataset (Indonesia, 2025) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.21838598

Song, Y. W. (G.), Lim, H. S., & Oh, J. (2021). “We think you may like this”: An investigation of electronic commerce personalization for privacy-conscious consumers. Psychology & Marketing, 38(10), 1723–1740. https://doi.org/10.1002/mar.21501

PwC Indonesia. (2026, September 3). Masyarakat Indonesia semakin mengandalkan teknologi, sumber terpercaya dan komunitas dalam menjalani gaya hidup yang lebih sehat, menurut survei PwC [Indonesians increasingly rely on technology, trusted sources, and community in pursuing healthier lifestyles, according to a PwC survey]. https://www.pwc.com/id/en/media-centre/press-release/2026/indonesian/voice-of-the-consumer-survey-2026.html

Republik Indonesia. (2022). Undang-Undang Republik Indonesia Nomor 27 Tahun 2022 tentang Pelindungan Data Pribadi [Law of the Republic of Indonesia Number 27 of 2022 on Personal Data Protection]. https://peraturan.bpk.go.id/Home/Download/224884/UU%20Nomor%2027%20Tahun%202022.pdf