Recommendation Algorithms and Consumer Choice: Investigating the Effects of Personalization and Choice Overload

Recommendation algorithms determine how many options consumers see and how well those options fit them, and the two design choices interact with a well-documented feature of consumer psychology: beyond a modest number of options, additional choice reduces satisfaction and raises the likelihood of deferring the decision. This paper models choice satisfaction and deferral as functions of recommendation set size and personalisation, with personalisation raising the fit of the best option and lowering the cognitive cost of evaluating the set. In a simulation across set sizes from 3 to 48 and three personalisation levels, satisfaction falls and deferral rises with set size at every personalisation level, but personalisation shifts both curves: without personalisation, deferral rises from 18% at three options to 51% at 48, whereas with high personalisation it rises only from near zero to 13%, and satisfaction at 48 personalised options (0.71) exceeds satisfaction at three unpersonalised ones (0.45). The paper argues that personalisation does not eliminate choice overload but raises the set size at which it bites, that the optimal set size depends on personalisation quality, and that platforms should treat set size as a tuned parameter rather than a default. Design and research implications are derived.