Most Behavior Change Doesn't Work. Here's What Does.

Introducing BehavioralStrategy.com

Most behavior change techniques don’t work.

I figured this out in 2011, at the start of my applied behavioral science career. The results of most behavior change techniques were disappointing. If someone wasn’t doing a behavior on day 1, they weren’t doing it on day 10. It didn’t matter which techniques you applied.

I realized pretty quickly that certain behaviors are a natural fit for certain people, and others aren’t. When you matched the right user group with the right behavior, things worked. When you didn’t, nothing could save you. Not more push notifications. Not gamification. Not cleverer nudges. Nothing.

I called this Behavioral Strategy, because the matching has to happen during the strategy and planning phase, before a single line of code gets written. If you get the strategy wrong, nothing downstream matters.

Since then, the evidence has caught up. Dramatically.

The Open Science Collaboration published their landmark study in Science in 2015 and found that only 36% of psychology findings replicated successfully. Social psychology was worse: roughly 25%. Effect sizes that did replicate were about half as large as originally reported. Many of the field’s most celebrated ideas (social priming, ego depletion, power posing) failed to replicate.

The nudge literature tells the same story. A review of 126 randomized controlled trials covering 23 million people found that real-world nudge interventions produce an average effect of about 1.4 percentage points. That’s roughly one-sixth the size of what gets reported in academic journals. A 2025 meta-analysis of 1,638 nudge studies found that after correcting for publication bias, the aggregate effect drops to essentially zero.

But individual differences research held up.

Personality traits replicate robustly across samples, cultures, and decades. The Big Five structure has been confirmed in dozens of countries. Heritability estimates for personality cluster around 40% to 50% in standard twin studies, and higher with more rigorous measurement methods. Test-retest stability is high and increases with age. When you aggregate behavior across contexts, trait-behavior correlations routinely reach the .40 to .60 range.

The most reliable finding in the behavioral sciences is this: people have stable, unique preferences and predispositions. Go against those predispositions, and you will fail. Work with them, and you have a real shot.

That is the foundation of Behavioral Strategy.

Over the past 10+ years, I’ve worked to codify this into frameworks, methods, and practical tools: the Four-Fit Hierarchy (Problem Market Fit, then Behavior Market Fit, then Solution Market Fit, then Product Market Fit), the DRIVE Framework (a five-phase methodology for getting the strategy right before you build), and Behavior Matching (a structured process for selecting the highest-fit behavior for a given audience and context).

I just published the first version of all of this at www.behavioralstrategy.com.

I plan to keep building this resource. If you design products, lead teams, or want to change your own behavior, start there.