New power-intensive industries and electrification of existing industrial processes put pressure on electric distribution grids. To evaluate whether flexibility measures such as procuring congestion management services can be considered to defer grid reinforcements, distribution system operators (DSOs) first need to characterize the need for flexibility in their grids. While characterization and modelling of flexibility resources have been intensively studied over the last years, characterizing the need for flexibility from the DSO’s perspective has received much less attention. This paper proposes a methodology for quantitatively characterizing the need for flexibility in distribution grids based on time series load demand data. The methodology and its application to active distribution grid planning is demonstrated for a real, industrial grid area in Norway with a varied set of industries and plans for electrification. The benefits were demonstrated by comparing probabilistic metrics for flexibility needs with the traditional methods currently used for grid planning by DSOs, which provide much more limited information.