Abstract We investigate how H ii region temperature structure assumptions affect “direct-method” spatially-resolved metallicity observations using multispecies auroral lines in a galaxy from the SAMI Galaxy Survey. SAMI609396B, at redshift z = 0.018, is a low-mass galaxy in a minor merger with intense star formation, analogous to conditions at high redshifts. We use three methods to derive direct metallicities and compare with strong-line diagnostics. The spatial metallicity trends show significant differences among the three direct methods. Our first method is based on the commonly used electron temperature Te([O iii]) from the [O iii]λ4363 auroral line and a traditional Te([O ii]) – Te([O iii]) calibration. The second method applies a recent empirical correction to the O+ abundance from the [O iii]/[O ii] strong-line ratio. The third method infers the Te([O ii]) from the [S ii]λλ4069,76 auroral lines. The first method favours a positive metallicity gradient along SAMI609396B, whereas the second and third methods yield flattened gradients. Strong-line diagnostics produce mostly flat gradients, albeit with unquantified contamination from shocked regions. We conclude that overlooked assumptions about the internal temperature structure of H ii regions in the direct method can lead to large discrepancies in metallicity gradient studies. Our detailed analysis of SAMI609396B underlines that high-accuracy metallicity gradient measurements require a wide array of emission lines and improved spatial resolutions in order to properly constrain excitation sources, physical conditions, and temperature structures of the emitting gas. Integral-field spectroscopic studies with future facilities such as JWST/NIRSpec and ground-based ELTs will be crucial in minimising systematic effects on measured gradients in distant galaxies.