Understanding the function of poly(A) tails has been challenging because of difficulties in measuring their lengths in a genome-wide manner. Bartel and colleagues now report a new high-throughput sequencing method, called PAL-seq ('poly(A)-tail profiling by sequencing'), that accurately measures individual poly(A) tails, irrespective of their length. By measuring the poly(A)-tail lengths of millions of individual RNAs isolated from vertebrate, fly, plant and yeast cells, the authors found that poly(A)-tail lengths were conserved among orthologous mRNAs, and that mRNA encoding ribosomal proteins and other 'housekeeping' proteins tended to have shorter tails. In early embryos of both zebrafish and Xenopus laevis, poly(A)-tail length was strongly correlated with translational efficiency, confirming and extending previous reports. However, this strong coupling largely disappeared during gastrulation, suggesting that translational control undergoes a mechanistic change that uncouples translational efficiency from poly(A)-tail length during development. The weak or negative correlation between tail length and translational efficiency in yeast and mammalian cells was confirmed by the observed lack of intragenic coupling between poly(A)-tail length and translational efficiency. The authors also analyzed miRNA-mediated poly(A)-tail shortening of target mRNAs during early zebrafish development. Interestingly, the developmental switch in translational control coincided with a switch in the predominant outcome of miRNA-mediated deadenylation, which changes from translational repression to mRNA destabilization during embryonic development. Overall, these findings reveal a transient relationship between poly(A)-tail length and translation in early embryos and provide a compelling explanation for why the ultimate effect of miRNAs on gene expression changes during development. (Nature, doi:10.1038/nature13007, published online 29 January 2014)