Understanding late Holocene to present relative sea-level changes at centennial or sub-centennial scales requires geological records that dovetail with the instrumental era. Salt marsh sediments are one of the most reliable geological tide gauges. In this paper we review the methodological and technical advances that promoted research on ‘high resolution’ late Holocene sea-level change. We work through an example to demonstrate different pathways to quantitative reconstructions of relative sea level based on salt marsh sediments. We demonstrate that any reconstruction is in part a result of the environment from which the record is taken, the modern dataset used to calibrate the fossil changes, statistical assumptions behind calibrating microfossil assemblages and choices made by the researchers. With the error term of typical transfer function models ~10–15% of the tidal range, micro-tidal environments should produce the most precise sea-level reconstructions. Sampled elevation range of the modern dataset also has a strong influence on model predictive ability. Model-specific errors may under represent total uncertainty which comes from field practices, sedimentary environment, palaeo-tidal changes and sediment compaction as well as statistical uncertainties. Geological tide gauges require a detailed chronology but we must be certain that apparent relative sea-level fluctuations are not simply a consequence of an age–depth model. We make six suggestions to aid the development and interpretation of geological tide gauge records.