Take regression / importance outputs and turn them into a simple “what if we improve this driver?” simulator.
All values are normalised to a Base Index = 100 for easy storytelling.
1. Paste driver importance data
Expected format (one row per driver): Driver, CurrentScore, MinScore, MaxScore, Importance.
“Importance” can be a regression coefficient, Shapley weight, or normalised importance score (all positive).
Example: Scores are on a 1–5 scale (mean ratings). Importance is how strongly each driver predicts your KPI (e.g., Overall Satisfaction or NPS).
2. Adjust drivers & see predicted impact
Base scenario is normalised to Index 100. Moving sliders changes the predicted KPI Index.
Base Index
100.0
New Index
100.0
Index Uplift
+0.0
Model logic: We create a simple linear KPI score = Σ(Importance × Score). Base scenario is rescaled to Index 100.
Any changes in driver scores move the KPI Index proportionally. This is meant for scenario directionality, not for exact forecasting.
Driver
Importance
Base score
Min–Max
New score (use slider)
Δ score
Tip: Start with realistic shifts (e.g., +0.2 to +0.5 on a 5-point scale) to avoid over-promising impact.
3. Visualise impact
Index impact and contribution of each driver to overall uplift.
KPI Index – Base vs Scenario
Driver contributions to uplift
The contribution chart shows how much of the total uplift is coming from each driver, given the changes you made.
If some drivers were moved down, they will show as negative contributions.
4. Scenario details by driver
Base vs new scores and each driver’s share of uplift.
Driver
Importance
Base score
New score
Δ score
Base contrib.
New contrib.
Uplift contrib.
Share of uplift %
“Contribution” here is Importance × Score (arbitrary units). Shares of uplift are based on the change in these contributions.
5. Quick interpretation
How to talk about this scenario with stakeholders.
Recommended language: “Indicative impact model based on regression / importance analysis. Numbers are directional and not a precise forecast.”
