Pareto Chart for Query Distribution
Chatbot traffic almost never spreads evenly across intents. A handful of common requests dominate, while a long tail of rare intents barely registers. This Pareto chart makes that "80/20" pattern visible so teams know where to focus their evaluation and optimization effort.
Interactive Demo
You can embed this MicroSim in your own page with the following iframe:
1 | |
Overview
A Pareto chart combines two views of the same data:
- Blue bars (left axis) show the query count for each intent, ordered from most to least frequent. A darker-to-lighter blue gradient reinforces the ranking, and the aggregated long tail is shown in gray.
- Red line (right axis) shows the cumulative percentage of all queries as you move left to right. It rises steeply, then flattens.
A dashed purple line marks the 80% threshold, and a vertical line plus a shaded green focus zone highlight the small set of intents (the "critical 20%") that together account for roughly 84% of traffic. In this dataset the 80% mark is crossed at the fifth intent.
Hover any column to see the exact count, that intent's share, and the running cumulative percentage.
Lesson Plan
- Read the chart: Have students identify how many intents are needed to reach 80% of traffic, and which intents fall in the long tail.
- Interpret: Ask why the cumulative line flattens. What does a flat segment tell you about the value of improving those intents?
- Prioritize: Given limited engineering time, which intents should the team optimize first, and why?
- Apply: Have students sketch what the chart would look like for a bot whose traffic is evenly distributed, and discuss why that is rare.