The synthesised consideration set: brand inclusion and concentration in large language model recommendations across five industries and three systems

Żatuchin, Dmitrij
Journal of Marketing Analytics · Springer Nature · 2026-10-31 · DOI: 10.1057/s41270-026-00562-3

Abstract

AI-generated supplier lists provide observable inputs to buyer search, but brand mentions do not establish consumer consideration or choice. Drawing on consideration-set and decision-aid research, this exploratory study defines the synthesised consideration set as the brands supplied in an answer. Four research questions examine inclusion, concentration, co-mention and cross-system agreement across five industries, 50 brands and 250 queries, repeated five times on three systems in February and September 2026 (3,614 and 3,750 scored answers). Query-level repetition coexists with dispersed average inclusion: February’s mean industry Gini is 0.30, while 204 queries have a brand named in at least 80% of answers. Most apparent coverage gaps reflect the sampled brand frame. The partially pre-specified September replication retains strong cross-date inclusion-share correlation (Spearman 0.994) and cross-system agreement of 60.8% versus February’s 57.2%. Fixed margins do not explain every co-mention pattern: 31 ordered pairs depart from September’s null. Model-adjudicated extraction checks were reviewed by the author. The contribution to marketing analytics is a set of denominator-explicit diagnostics of supplied brand information. Proposed managerial uses concern audit coverage and prioritising further tests; consumer effects, causal mechanisms and commercial returns remain unmeasured.

Read at publisher →

Similar articles