The hardest part of analytics is not finding insights—it is getting people to act on them.
The Insight-Action Gap
Every analyst has experienced this: you spend weeks on a rigorous analysis, present compelling findings, get nods of agreement… and then nothing changes. The insight dies in a slide deck.
I have watched this happen more times than I care to admit—including with my own work. Why does this happen?
1. Insights Without Context
Data does not speak for itself. An insight like “retention dropped 5%” means nothing without context:
- Is this normal seasonal variation?
- How does it compare to competitors?
- What is the business impact in dollars?
Fix: Always frame insights in terms stakeholders care about—usually revenue, cost, or strategic goals.
2. No Clear Owner
“Someone should look into this” is where insights go to die. If there is no specific person accountable for acting on a finding, it will not happen.
Fix: End every analysis with explicit recommendations and named owners. I have started doing this even when it feels awkward.
3. Too Many Insights
Presenting 20 findings dilutes attention. Executives have limited bandwidth. They cannot act on everything.
Fix: Ruthlessly prioritize. Lead with the 1-3 insights that matter most. Put the rest in an appendix.
The ARIA Framework
I use a simple framework for insights that drive action:
- Actionable: Can someone actually do something about this?
- Relevant: Does it connect to current business priorities?
- Impactful: Is the potential value significant?
- Assigned: Is there a clear owner and timeline?
If an insight fails any of these tests, it is not ready to present.
Building Influence Over Time
Driving action is not just about individual analyses—it is about building credibility:
- Start small: Prove value with quick wins before tackling big questions
- Follow up: Track whether your recommendations were implemented and what happened
- Be honest about uncertainty: Stakeholders trust analysts who acknowledge limitations
- Learn the business: The best analysts understand operations, not just data
The goal is not to be right. It is to make the business better. Sometimes that means accepting that a perfect analysis matters less than a good-enough analysis that actually gets used.