Previous research efforts have focused on developing prospective life cycle inventory databases that build upon projections from integrated assessment models but were limited to attributional system models. A novel approach is required to construct consequential LCI databases that can be applied consistently on a large scale. To this end, the heuristic approach from Bo Weidema was selected as a basis for this study. This approach has been validated with historical data and was adapted in this study to identify the marginal suppliers in a prospective context. The different steps within the approach were analyzed, and alternative techniques for each step within the heuristic method were proposed. The techniques were tested on the future electricity sector using projections from two integrated assessment models (IMAGE and REMIND). Results show the sensitivity of results on the modelling technique selected in each step. The most sensitive step is the selection of the time interval, with even small changes resulting in a noticeable difference. In addition, the results also showed a substantial difference between the projections of the two models. The relevance and goals of the alternative techniques for each step were discussed to guide users in forming the heuristic method for their study.