Conference Agenda

Continuous improvement programs are a cornerstone of operational excellence in manufacturing and logistics. Firms rely on structured idea management systems to channel employee suggestions for process improvements into tangible efficiency gains. Yet realizing the operational value of these ideas depends not only on selecting the right ones, but on doing so promptly. Delayed evaluations erode the value of approved improvements and, as research on procedural justice suggests, undermine the cooperative behavior of employees who perceive the process as slow or unfair. Despite this practical importance, prior research has focused almost exclusively on the outcome of idea evaluation — which ideas get selected — leaving the process largely unexplored. This paper asks: what factors drive idea evaluation time in operational improvement systems, and how can we explain their impact?

We draw on longitudinal data from the internal idea management system of a large European manufacturer in the mobility industry. The dataset covers more than 1.2 million evaluation activities from over 30,000 unique evaluators for roughly 240,000 submitted kaizen ideas spanning 14 years (2004–2018), with precise timestamps for each evaluation step. Ideas range from small operational upgrades (e.g., tool mounting improvements) to significant process changes (e.g., restructuring manufacturing workflows).

Given that theory on evaluation speed is nascent, we follow a two-stage approach. First, we apply algorithm-supported induction: we train a gradient boosting model (XGBoost) and use Shapley values (SHAP) to surface patterns in the data, which we then embed in existing theory to formulate hypotheses. Second, we test these hypotheses on a held-out sample using a two-way fixed-effects OLS regression.

Results reveal several robust drivers of evaluation time. Evaluations take roughly one-third longer when evaluator and ideator share the same organizational unit, pointing to coordination overhead or heightened local scrutiny. Involvement of higher-hierarchy evaluators in prior steps cuts evaluation time by around one-fifth, suggesting senior participation resolves operational ambiguity or raises process priority. A higher average duration of preceding activities increases the focal evaluation time — consistent with collective shirking — while longer prior-evaluator time reduces it, suggesting effort substitution. Evaluator workload increases delays; ideator experience and recent dual-role activity (having submitted an idea oneself) reduce them.

The findings offer actionable levers for operations managers seeking to accelerate improvement cycles, reduce procedural drag, and sustain employee engagement in continuous improvement programs.