Pareto Principle (80/20)
A minority of inputs often drives a majority of outcomes. Find the vital few, focus there first.
Author
Vilfredo Pareto (1896); generalised by Juran as the “vital few vs trivial many”
Model type
About
The Pareto Principle is a practical heuristic drawn from skewed (often heavy-tailed) distributions: results are unevenly concentrated. It’s not a law of nature and not always 80/20, but the pattern is common—few products → most revenue, few defects → most failures, few tasks → most impact. The move is to identify concentration, act on the top contributors, and design a sane policy for the long tail.
How it works
Skewed contribution – rank items by impact (descending); cumulative impact rises steeply then flattens.
80/20 is a placeholder – real cut-points might be 70/30, 90/10, 65/35—use your data.
Visuals –
- Pareto chart: bars (impact per item) + line (cumulative %).
- Lorenz curve: share of items vs share of outcome; bowing indicates inequality.
Why it appears – compounding effects, network/exposure differences, constraints, positive feedback, and variance.
Action logic – handle the vital few with disproportionate attention; standardise or batch the useful many; cap or drop the trivial tail.
Use-cases
Revenue & customers – top cohorts, SKUs, or channels drive most contribution.
Quality & reliability – few root causes explain most incidents/defects.
Product & UX – a handful of journeys/features create most value; simplify around them.
Growth & content – few pages/creators pull most traffic; optimise those first.
Inventory & procurement – ABC classification for stocking and vendor attention.
Personal effectiveness – identify tasks/meetings that actually move outcomes.
Step-by-step
Define the outcome – revenue, failures, time, adoption, cost.
Collect and rank – list items (SKUs, causes, tasks) with measured impact; sort high → low.
Plot a Pareto chart – add cumulative %; mark where the curve bends (natural threshold).
Choose cut-offs – e.g., A (top ~10–20 percent of items) → high attention; B (next ~30 percent) → standard playbook; C (rest) → automate, cap, or prune.
Act on the vital few – bespoke fixes, premium SLAs, targeted experiments, root-cause elimination.
Set tail policy – minimum service levels, self-serve only, or discontinuation.
Review cadence – re-run monthly/quarterly; watch for drift as mix and seasonality change.
Pitfalls & Cautions
Worshipping 80/20 – use your distribution; don’t force the ratio.
Mean vs margin – optimise incremental ROI, not just biggest bars historically.
Tail blindness – small items may be strategic (emerging segments, reliability hotspots).
Goodhart risk – if you pay only on the top slice, gaming and neglect follow; add guardrails.
Stale rankings – today’s A-items can decay; keep the chart live.
Correlation ≠ cause – big bars show where to look, not why it happens.
Related Mental Models
Click below to learn other mental models
The Idea Maze
Before building, map the space: the key forks, dead ends and dependencies—so you can choose a promising path and run smarter tests.
The Map is not the Territory
All models are simplifications. Don’t confuse the representation (map, metric, plan, narrative) with the thing itself—and update the map when facts change.
Bottlenecks
Flow moves at the pace of its constraint—improve the bottleneck to improve the whole.