This study examined the key factors influencing the retention of administrative staff in Kazakhstani higher education institutions (HEIs), a workforce whose turnover has received little dedicated research attention despite its cost to institutional stability and service quality. Methodology – A qualitative research design was adopted, using semi-structured interviews with ten administrative staff (seven women, three men; 1–13 years of experience) purposively sampled from one private and one autonomous university in Astana and Almaty. Data were analysed using inductive thematic analysis, guided by Herzberg's two-factor theory. Originality/value – Administrative staff retention in Kazakhstan remains an under-researched area; the study contributes context-specific, practice-oriented evidence to a thin regional literature base and tests the applicability of Herzberg's framework in a non-Western higher education setting. Findings – Retention was shaped by salary, workload, career opportunities, corporate culture, leadership style, team relationships, benefits, professional development, and job stability. Career growth emerged as the strongest driver, consistent with Herzberg's motivators, while salary functioned as both a hygiene factor and, at times, a direct motivator — a nuance not fully captured by the original theory. Younger staff prioritised pay and career growth; more experienced staff prioritised comfort and stability. The autonomous university offered stronger formal benefits but a more bureaucratic culture; the private university offered flexibility and informal relationships but weaker structured support. The findings suggest that retention strategies in Kazakhstani HEIs should be tailored by career stage and institution type, with HR departments taking a more continuous, proactive role across the employee lifecycle rather than engaging mainly at recruitment and exit.

administrative staff; staff retention; turnover intention; higher education institutions; Kazakhstan; Herzberg's two-factor theory; qualitative research