Oil palm improvement is frequently discussed as a search for progressively higher fresh fruit bunch (FFB) or crude palm oil (CPO) yield. Such a single-trait perspective is increasingly inadequate because economically successful planting materials must simultaneously express high productivity, early bearing, favourable bunch architecture, manageable vegetative growth, efficient oil recovery, environmental stability, and resistance to major biotic and abiotic stresses. This qualitative literature review integrates contemporary research published primarily during 2020–2026 with an exploratory statistical analysis of 54 Indonesian oil palm varieties released between 1984 and 2017. Catalogue-level analysis reveals a strong negative relationship between bunch number and average bunch weight (Spearman ρ = −0.885, p < 0.001), a positive association between FFB productivity and reported oil output (ρ = 0.619, p < 0.001), progressively earlier harvesting in newer releases (ρ = −0.566 with release year), and higher reported CPO/oil productivity among later generations. However, environmental contrasts within individual genetic materials demonstrate that maximum recorded yield cannot be interpreted independently of growing environment. Recent genomic selection, GWAS, QTL mapping, transcriptomics, metabolomics, high-throughput phenotyping, and disease-resistance research collectively support a shift toward multi-trait, environment-specific breeding. The review proposes a selection architecture centred on robust lifetime oil productivity rather than maximum short-term yield, integrating productivity, precocity, harvestability, extraction efficiency, genotype × environment stability, climate resilience, disease resistance, and deployment economics.