An excess of flux (i.e. a bump) in the early light curves of type Ia supernovae has been observed in a handful of cases. Multiple scenarios have been proposed to explain this excess flux. Recently, it has been shown that for at least one object (SN 2018oh) the excess emission observed could be the result of a large amount of 56Ni in the outer ejecta (∼0.03 M⊙). We present a series of model light curves and spectra for ejecta profiles containing 56Ni shells of varying masses (0.01, 0.02, 0.03, and 0.04 M⊙) and widths. We find that even for our lowest mass 56Ni shell, an increase of >2 magnitudes is produced in the bolometric light curve at one day after explosion relative to models without a 56Ni shell. We show that the colour evolution of models with a 56Ni shell differs significantly from those without and shows a colour inversion similar to some double-detonation explosion models. Furthermore, spectra of our 56Ni shell models show that strong suppression of flux between ∼3700–4000 Å close to maximum light appears to be a generic feature for this class of model. Comparing our models to observations of SNe 2017cbv and 2018oh, we show that a 56Ni shell of 0.02–0.04 M⊙ can match shapes of the early optical light curve bumps, but the colour and spectral evolution are in disagreement. Our models also predict a strong UV bump that is not observed. This would indicate that an alternative origin for the flux excess is necessary. In addition, based on existing explosion scenarios, producing such a 56Ni shell in the outer ejecta as required to match the light curve shape, without the presence of additional short-lived radioactive material, may prove challenging. Given that only a small amount of 56Ni in the outer ejecta is required to produce a bump in the light curve, such non-monotonically decreasing 56Ni distributions in the outer ejecta must be rare, if they were to occur at all.