PM Optimizer
Set up experiments that vary your model's settings, find what matters, and search for the best configuration.
- PM OptimizerPM Optimizer runs experiments that vary your model's settings, measure the effect on your results, and search for the best configuration.
- Define an experimentThe Define step in PM Optimizer: name the experiment, set each response's goal and weight, add limits, and read the run footer before you start.
- FactorsFactors are the settings PM Optimizer may change: ranges, option lists, on-or-off choices, model changes, and logic, across everything in your model.
- ResponsesResponses are what a PM Optimizer experiment measures: whole-process, per-step, per-entity-type, resource, and inventory measures, plus your own data.
- DesignsThe design controls how a PM Optimizer experiment explores the combinations: Full Factorial or the Smart Optimizer, plus screening to find what matters first.
- Running an experimentThe Run step in PM Optimizer: a live dashboard with progress, a best-so-far chart, paired comparison, stopping, and what happens at completion.
- Reading the resultsThe Results step in PM Optimizer: the ranked table with a baseline row, statuses, main effects, comparison with confidence ranges, and saving the winner.
- Smart OptimizerThe Smart Optimizer in PM Optimizer: a search that learns from every run, with a run budget, auto-stop, automatic replications, and a multi-goal frontier.
- Optimize ThisOptimize This turns a report finding into a ready-made experiment: it picks the levers, measures today's output, screens the factors, and sets up the search.