Purchase Intent Fine-Tuning Simulator
Published model · Gradual Distributional Shifts · ← All-or-Nothing Simulator
What if the mix of package reactions had been slightly different — a few more "very relevant" ratings, a few fewer "not premium" responses?
This simulator runs on the published model — six package-reaction questions chosen for the clearest, most interpretable story about purchase intent.
Shift the mix of package reactions gradually — moving 15% of neutral ratings to positive rather than moving every respondent at once.
How to Use This Tool
Shift response distributions continuously and see how purchase intent would have changed
How to Use This Tool
Shift response distributions continuously and see how purchase intent would have changed
1. Try a preset or drag sliders
Use the preset buttons for a quick start, or drag the sliders to change the mix of responses for any variable. The colored bar under each variable's header shows its model leverage — how much the predicted outcome can swing based on that variable alone. Each variable's percentages must sum to 100.
2. Run the scenario
Click Run scenario to send your slider settings through the published logistic regression model. The result shows the projected top-2 purchase intent and a 95% confidence interval. The How certain is this result? chart shows 10,000 simulated outcomes so you can see how much the baseline and your scenario overlap.
3. Explore each factor
Click Explore Each Factor after a run to see a full per-variable sensitivity breakdown. Each card shows the predicted purchase intent if that one variable's distribution were set to 100% at each level, holding your other sliders fixed. Combined best/worst cards show the result of every variable's best or worst level at once.
4. Review past scenarios
Every run is saved to the Saved Scenarios drawer at the bottom of the page. Each card shows which variables you shifted from baseline and the predicted outcome. Click Load to restore a past scenario's slider settings, or Remove to drop it.
5. Shift distributions gradually
Unlike the All-or-Nothing simulator, sliders let you move just some respondents from one answer to another — for example, shifting 10% of "Disagree" ratings to "Neither agree nor disagree." This tests realistic, incremental shifts rather than extreme all-or-nothing scenarios. Make sure each variable's totals still sum to 100%.
6. Reset and iterate
Click Reset to baseline to return all sliders to the empirical concept-test distributions and clear the results. The baseline of 54.7% is the model's estimate of top-2 purchase intent in the unmodified concept-test sample. Try different combinations to see which package factors matter most — or least.
Loading simulator…
Couldn't load the simulator
Set the distribution
The number on each card is the variable's swing on this page — how far the prediction would move if you concentrated all responses in this variable's lowest-leverage level versus its highest-leverage level, using the real values within each level. A bigger swing means redistributing this variable matters more.
Note: this page doesn't pin everyone to one specific value. It shifts what kind of responses show up in the model's average — emphasizing some, de-emphasizing others — so the answers stay grounded in situations the data actually contains. That's why swings here are smaller than on the All-or-Nothing page, where every respondent is set to the same extreme value.
Predicted outcome
How did you change the survey?
Each pair of bars compares the actual distribution (green) to the distribution under your scenario (red where you changed it). This is what you changed — not the predicted impact, just the input.
How certain is this result?
Every prediction has wiggle room — these histograms show how much. The green bars are the plausible answers for the baseline; the red bars are the plausible answers for your scenario. Where the colors overlap, the two answers are close enough that the model can't cleanly tell them apart.
Calibration: predicted vs. observed
How closely the model's predicted probabilities track the observed outcome rates, binned by predicted decile. Bubble size shows respondents per bin. The dashed diagonal is perfect calibration; the blue scenario line is your run; the gray line (when present) is the unmodified base model for comparison.
Explore each factor
For each variable, see how the predicted outcome would change if you switched that one distribution to 100% at each level — holding all your other selections fixed. Reveals which individual levels have the most leverage given your current scenario.