Skills CX operations
Demand driver attribution
Leadership sees a spike and asks why. Support sees the spike and often answers with the last thing that changed — a release, a email, a price update. Coincidence is not attribution.
npx rulebase-skills install cx-demand-driver-attributionWhen to use it
Reach for this when someone says any of these — they are the phrases the skill itself triggers on:
- “what caused the volume spike”
How it works
The method, in the order the skill runs it. The full procedure — tables, worked examples and the edge cases — is in the skill itself.
Start with a calendar, not a model
Maintain one demand calendar owned by ops, fed by product, marketing, billing and engineering.
Unit of analysis
Attribute at the granularity decisions use: Contact driver tags, Product area / queue / SKU, Customer segment, Interval.
Difference-in-differences mindset
You rarely have a randomised experiment. DiD is the disciplined informal frame: Treated, Control, Before / after, Difference of differences.
Confounders that fake attribution
Check these before presenting a driver: Seasonality and DOW, Overlapping events, Routing and tagging changes, Deflection and channel shift, Repeat contact inflation, Internal backlog release.
What you can and cannot conclude
Can conclude (with evidence): Treated segment volume rose relative to control over the same window, Effect decay pattern matches mechanism.
Related skills
Free and open source, and vendor-neutral — it reads the conversations from whichever helpdesk you already run. Browse all 149 skills · connect your helpdesk over MCP · source on GitHub
