Purpose: Life cycle assessment (LCA) is being used increasingly in decision support situations. In actual cases, the sources of uncertainty are easily hidden in the complexity. Methods for taking uncertainty into account are recommended by LCA guidelines, but actual application remains rare. The aim of this study is to demonstrate the sources of uncertainty in a practical simple selection case wherein a customer makes a decision between beer and wine in a restaurant, considering the selected criteria and the given information. The uncertainty in LCA results is connected to the broader scope of decision analysis. Methods: Life cycle inventories were collected for beer and wine production from existing literature. The functional unit was chosen to be one serving of alcohol: beer or wine. For illustrative purposes, only the global warming potential indicator was included in the LCA through carbon footprint (CF). Probabilistic uncertainty analysis was applied to the CF system using Monte Carlo simulation. Water footprint was also roughly considered. In addition, three non-environmental indicators were included in the decision: weight control, price, and taste. The comparison between the two products was constructed as a multiple-criteria decision analytical problem. Results and discussion: The results indicated that beer had, on average, a higher CF value than wine did. However, the difference was not significant, and within the uncertainty range, also the opposite conclusion was possible. The ratio of wine to beer CF was dominated by the uncertainty in the NO emissions of wine production. When all of the decision criteria were included, the level of uncertainty prevented robust overall conclusions about preference for beer or wine. However, depending on the utility differences assigned to subjective indicators, there existed also cases wherein decisions could be made at a 10 % risk level regardless of high overall uncertainty. Conclusions: In many cases, the uncertainties of LCA are dwarfed by the overall uncertainty of the decision situation. However, as shown by our example, in many cases, reasonable decisions can be made in spite of high uncertainties. The uncertainties of single LCA indicators should be considered in relation to the decision-making problem, which depends on the uncertainty of LCA indicators but also significantly on the weighting of the indicators and the related uncertainty. Successful decision making depends on both the magnitude of uncertainty and the differences in expected utility value between alternatives. More attention should be paid to uncertainty analysis considering the weighting factors. [ABSTRACT FROM AUTHOR]