Conversions Matter / Frameworks
Better thinking.
Better decisions.
Practical Frameworks for Analytics, Experimentation and Responsible AI
Frameworks for making sense of evidence, using AI responsibly and improving the decisions we make.
These frameworks capture how I approach analytics, experimentation and the problems that sit between data and decisions. They draw on established thinking, lessons from other practitioners and nearly two decades of doing the work. I'm sharing them because our industry improves when we share what we've learned and challenge each other's assumptions.
Explore the frameworks
Different problems call for different tools. Each framework stands on its own, but they share a principle: make the reasoning visible, test the assumptions and learn from what happens.
Responsible AI
S.P.I.D.E.R.
A practical check for using AI responsibly, not just enthusiastically.
Sourced · Private · Inclusive · Disclosed · Evaluated · Reversible
Six principles to consider before deploying AI in research, analytics, experimentation or customer experiences. Designed to make responsible use a practical decision rather than an abstract aspiration.
Explore S.P.I.D.E.R. →Analytical reasoning
PEEP
Stop at the observation and you've only done part of the job.
Point · Evidence · Explanation · Prediction
An adaptation of the familiar Point, Evidence, Explanation structure, with Prediction added for analytics. A simple way to turn findings into explanations we can challenge and expectations we can check.
Explore PEEP →What these frameworks are, and aren't
I'm not claiming to have invented scientific reasoning, ethical AI or better decision-making. This is a synthesis of ideas that have influenced my work, adapted into methods I find useful in practice.
Credit the foundations and be clear about what's been adapted.
Use examples and evidence, including uncomfortable or inconclusive results.
If a framework doesn't survive contact with practice, improve it.
Why share them?
Analytics and experimentation should help people make better decisions. Too often, we get caught up in producing more reports, running more tests or adopting another tool without asking whether we're solving the right problem.
These frameworks are my attempt to make the thinking more explicit and useful. Take what helps, question what doesn't and improve it if you can.
