Go beyond correlation to answer the question that matters most: "If I improve this driver, what return will I get?"
Available via self-service platform and Alchemist Agent. Upload your data, get a boardroom-ready strategic report in seconds – no data scientist required.
A driver can be highly correlated with satisfaction but offer low returns if you're already performing well. Traditional analysis can't distinguish between "critical" and "already excellent."
A driver might show modest correlation but offer exceptional ROI because current performance is low. Traditional analysis would deprioritise this high-value opportunity.
Not all important themes are created equal. Some themes lift the score when improved; others, the Protect and Maintain themes, damage the score if they slip but do not move it higher if improved further.
Catalyst combines importance, impact, and performance to create a complete investment picture
"How well does this driver predict the outcome?"
Machine learning captures non-linear relationships that traditional correlation misses. Output: Importance Score (0-100).
"How much outcome change results from improving this driver?"
Simulates improving each driver while holding others constant. Reveals which drivers have steep slopes vs. diminishing returns. Output: Impact Index (0-100).
"Where do you currently stand on each driver?"
Low performance + High impact = Immediate opportunity. High performance + High importance = Protect but don't over-invest. Output: Performance Score (0-10).
Four base quadrants from the importance-and-impact view, plus two further reads that the Performance layer surfaces
These themes drive the outcome and there is room to move the score, so act here first.
There is room to move the score on these themes but they do not strongly drive the outcome, so investigate before investing.
These themes drive the outcome but there is little room to move the score, so safeguard current performance rather than invest heavily.
You are already performing well on themes that drive the outcome, so further investment yields diminishing returns. The Performance layer flags where additional spend would not move the score.
These themes neither drive the outcome nor have room to move, so deprioritise.
cxCatalyst is the productised v1 of the Catalyst framework, focused on the highest-volume use case: NPS plus verbatims. Upload your file, choose your splits, and a branded strategic report lands in your inbox. The four base classifications, the driver table, the confidence verdict.
Most CX programs rely on stated importance to prioritize action. But what if the attributes customers rate as important aren't actually the ones that drive their behaviour? We built a synthetic retail banking dataset to demonstrate how the Catalyst Framework reveals hidden priorities.
→ Invest heavily. Customers understate its importance, but it's the #1 driver of NPS. Every 1-point improvement generates 2.1× more value than Mobile App.
→ High-impact opportunity. Low current performance + high behavioural impact = immediate ROI. Service recovery is a loyalty multiplier.
→ Maintain, don't over-invest. Traditional analysis flagged this as #1 priority. But you're already excellent – further investment yields diminishing returns.
→ Hold the line. High importance, low impact. If fees clarity slips it damages the score, but improving it further does not move the score higher. Maintain current levels.
To demonstrate the Catalyst Framework, we used a synthetic dataset designed to reflect realistic South African retail banking customer experience dynamics. This is not actual FNB customer data – it's a simulation built to illustrate how the framework reveals hidden priorities.
The synthetic dataset was constructed using:
This approach mirrors how the Catalyst Framework would be deployed on actual client data. Knowsis offers synthetic data generation services for analytical prototyping, concept validation, and scenario planning – allowing you to test frameworks before committing to full research investment.
See how Catalyst sits inside the full methodology that turns segmentation into commercially viable personas, with the Intrum European Consumer Payment Report case study as the proof point.
Get The Predictive Persona Playbook →19 pages. Free. Delivered to your inbox.
Stop guessing where to invest. Get clear, defensible, ROI-focused priorities that drive measurable business outcomes.
Start Driver Analysis