From 80/20 Analysis to Action

Why analysis stalls after the presentation

A strong 80/20 analysis can produce dozens of observations. That breadth feels valuable, but it creates a management problem: every observation can begin to look like a priority.

The handoff fails when the team records findings without making decisions, creates initiatives without sizing value, or tracks activity without returning to the original economic case.

The purpose of the next stage is to 80/20 the 80/20—to narrow the analytical output to the few changes whose economics and feasibility justify ownership.

Separate findings from opportunities

A finding describes what the data shows. An opportunity describes a change the organization might make because of that evidence.

• Finding: a defined customer group has low margin and high order frequency
• Opportunity: change the service, pricing, or ordering conditions for that group to improve contribution

The first statement is diagnostic. The second proposes a mechanism. Keep them separate so the team can challenge whether the proposed action actually follows from the evidence.

For every candidate, record:

• the observed pattern
• the affected customer, product, or process group
• the suspected operational or commercial cause
• the proposed change
• the measure expected to move
• key constraints and dependencies

Apply four decision filters

1. Materiality

Estimate the revenue, margin, cost, capacity, working capital, or complexity affected. Use a range when the data does not support a precise number. Small improvements can still matter, but they should not displace larger, more defensible opportunities without a reason.

2. Specificity

The change should be clear enough that an owner knows what will be different. “Improve the tail” is not a change. A defined policy, customer conversation, price condition, product decision, process redesign, or service standard is.

3. Controllability

Identify what the organization can actually influence. A pattern driven mainly by market demand, regulation, contract, or customer strategy may require a different response than an internally controlled process condition.

4. Measurability

Name the implementation measure and the business outcome measure. If neither can be observed within a useful timeframe, the team will struggle to manage the work or learn from it.

Size the economic effect

Build a simple bridge from the current condition to the expected condition. Make the major drivers visible: affected volume, price, margin, cost, frequency, or capacity. Avoid a single unexplained “opportunity value.”

Use scenarios when behavior is uncertain:

• a conservative case based on limited adoption or realization
• an expected case based on the best current assumptions
• an upside case only when its additional assumptions are explicit

Separate recurring economic effect from one-time cost. Record the time required to realize the effect and the operational risks that could offset it.

The purpose is not to guarantee a result. It is to make the decision logic reviewable.

Select the critical few

Compare candidates using a small number of transparent dimensions: materiality, confidence in the mechanism, effort, time to impact, strategic fit, and risk.

Do not let a composite score replace discussion. Scores can create false distance between opportunities and imply precision that the estimates do not support. Use them to structure a conversation, then document the actual decision.

For each selected opportunity, record why it was chosen and why other meaningful findings were deferred. A defer decision is useful; it prevents the same unselected idea from reappearing as an implied priority in every review.

Define ownership and milestones

An accountable owner needs authority to change the relevant condition, not merely responsibility for updating a tracker.

Define:

• one accountable owner
• the affected functions and decision rights
• the first observable milestone
• dependencies and approvals
• the business outcome measure
• the review cadence
• the condition that would stop or redesign the work

Milestones should represent evidence of change, not administrative activity. “New service policy approved” is stronger than “meeting completed.”

Review activity and economics together

Use two connected views in each review:

1. Implementation: decisions made, actions completed, risks, milestones, and next commitments
2. Outcome: the customer, product, margin, cost, or complexity measure that justified the initiative

Outcome measures may lag, so pair them with leading indicators. Do not allow the leading indicator to become the definition of success. The original business effect remains the reason the work was approved.

When results differ from the case, ask whether execution, timing, assumptions, external conditions, or the analytical diagnosis caused the gap. The answer should improve the next decision, not merely explain the last report.

Build a closed learning loop

An 80/20 implementation should change the next analysis. Refresh the underlying customer and product views, compare classification and profitability changes, and determine whether the selected actions altered the intended distribution.

This closes the loop:

1. analyze concentration and economics
2. select the critical few changes
3. implement with ownership and measures
4. review the business effect
5. use the new evidence to make the next decision

COMPASS organizes this as Analyze, Implement, and Manage so the business case does not disappear when the work moves from analysis into execution.

Frequently asked questions

How many initiatives should come from an 80/20 analysis?

There is no fixed number. Select only the changes whose value is material, mechanism is understood, ownership is clear, and outcome can be measured. The purpose is to reduce priorities.

What makes an 80/20 finding actionable?

An actionable finding identifies a defined group, a specific change, a credible economic mechanism, an accountable owner, and an observable measure.

How should results be reviewed?

Review both implementation progress and the business measure that justified the work. Activity alone does not show whether the expected economic effect occurred.