Goal Attainment and Confidence, Not Satisfaction, Drive Office Workers' Paid ChatGPT Subscriptions
An analysis of 349 office workers finds that satisfaction did not translate into actual use, while goal-congruent outcomes and self-efficacy drove use and subscription intention — published in the international journal MIP
Researchers at the HJ Institute of Technology and Management (HJITM) have published a study analyzing the factors that determine office workers' intention to subscribe to the paid version of ChatGPT (ChatGPT Plus) in Marketing Intelligence & Planning (MIP, SSCI), an international journal in marketing.
The researchers surveyed 349 Korean office workers who use ChatGPT for work and analyzed the data using structural equation modeling. They examined whether ChatGPT's system quality, information support, service quality, and perceived intelligence lead through satisfaction and 'goal-congruent outcomes' (the perception of having achieved results that match one's work goals) to use and subscription intention, and what role users' self-efficacy plays.
The analysis showed that system quality (β=0.332), perceived intelligence (β=0.320), and service quality (β=0.242) significantly affected satisfaction. Service quality (β=0.642) had the strongest effect on goal-congruent outcomes, and perceived intelligence (β=0.168) was also significant. Information support, however, had no significant effect on either satisfaction or goal-congruent outcomes.
Notably, the results diverge from traditional models of user behavior. Satisfaction had no significant effect on actual use (β=0.054), whereas goal-congruent outcomes significantly increased use (β=0.427). Users' self-efficacy, their belief that they can handle ChatGPT well, also substantially increased use (β=0.446), and use in turn led strongly to subscription intention (β=0.486). The interpretation is that what drives office workers to pay is not the feeling that the experience was good, but the outcome of having achieved their work goals and the confidence that they can use the tool well.
The study is notable as the first empirical examination of a new consumer behavior: paid subscriptions to AI chatbots. It suggests that AI service companies can increase subscription conversions not by stopping at satisfaction surveys, but by strengthening features and service quality that help users achieve their work goals and by raising users' confidence through training on how to use the tools.
HJITM continues its research on the business use of generative AI and user behavior.
Key findings
- System quality (β=0.332), perceived intelligence (β=0.320), service quality (β=0.242) → satisfaction
- Service quality (β=0.642), perceived intelligence (β=0.168) → goal-congruent outcomes; information support has no effect
- Satisfaction → use not significant (β=0.054); goal-congruent outcomes (β=0.427) and self-efficacy (β=0.446) → use
- Use → subscription intention (β=0.486) · N=349
- Paper
- Subscription intentions for ChatGPT plus: a look at user satisfaction and self-efficacy
- Authors
- Hyeon Jo
- Journal
- Marketing Intelligence & Planning (Emerald, SSCI), Vol. 42(6), pp. 1052–1073, 2024
- DOI
- 10.1108/MIP-08-2023-0411
- Cite
- Hyeon Jo (2024). Subscription intentions for ChatGPT plus: a look at user satisfaction and self-efficacy. Marketing Intelligence & Planning, 42(6), pp. 1052–1073.
This story could become your next paper
HJITM works with researchers from study design through surveys, analysis and submission to international journals in generative AI and technology management. Bring us your research question.