Concentrated leverage
Concentrated leverage is an approach to improvement work that starts from the minority of customers who actually pay back what it cost to acquire them, and asks what they're doing that the rest aren't. The signal lives in that minority, not in the cohort average, so lifting the average across the whole population washes out the very pattern that distinguishes them. The work is investigative before it's interventional: find what those customers are doing, then design the product so more users end up doing it.
Definition
The discipline of identifying the behavioral patterns that distinguish profitable customers from missed conversions, then designing to produce more of those patterns at scale.
Profitability-relevant behavior is concentrated
Profitability-relevant behaviors aren't uniformly distributed across a signup cohort. Profitable customers — the signups who reach the offering's anchored commercial outcome — are concentrated in a minority, and their patterns differ systematically from missed conversions. Conventional activation work treats the cohort as a single unit with an activation rate to lift, which dilutes the minority's distinctive signal into cohort noise.
Lifting the average can move the metric without amplifying the minority. Looser gating, friction reduction, or universal onboarding can shift the rate while leaving the profitable-customer minority untouched. In the homogenization case, the same moves erode the very signal that distinguishes the minority, by smoothing the cohort on dimensions that used to discriminate.
Forensic before interventional
The question shifts from "how do we move the rate" to "what are profitable customers doing that missed conversions aren't." Improvement begins by surfacing what's already producing economic gain in the cohort, then designs to produce more of it.
Multiplier analysis on the signup cohort is the primary forensic operation: candidate behaviors get ranked by how much they distinguish converters from baseline. The interventions that follow — path investment, detection tuning, affordance design — amplify what the analysis surfaced rather than operating on faith about what should work.
Growing the profitable minority
The minority isn't a static segment to be served exclusively. The 10% is a target population to grow. The 90% decomposes into two groups: recoverable missed conversions, whose intent the offering can serve but who weren't served by the activation path; and structurally unrecoverable signups, whose intent the offering doesn't serve at all. Concentrated leverage's primary work targets the first — designing infrastructure that helps more recoverable candidates complete the profitable-customer pattern.
The structurally unrecoverable subset isn't actively rejected. It falls through to OSFA like any user without a built path. The discipline is about where path investment goes — toward converting recoverable missed conversions, rather than chasing structurally unrecoverable signups into conversions that won't sustain.
Relation to outcome anchoring
Commercial-outcome anchoring fixes what the apparatus answers to; concentrated leverage is the discipline brought to the visible distribution under that anchor. The two are conceptually distinct (they fail differently when missed) and operationally entangled (anchoring is what makes the profitable-customer minority visible; concentrated leverage is what's done with the visibility). A team that anchors correctly but optimizes the cohort average misses the leverage. A team that practices concentrated leverage against the wrong anchor amplifies non-sustaining patterns. Both moves are required; neither substitutes for the other.
Example
A team's activation rate is 12%. Two routes are available.
The cohort-average route: simplify the signup flow, defer email verification, remove the workspace-naming step. The rate moves to 15%. New conversions look like average existing conversions — some sustain, some don't, in roughly the same proportions as before.
The concentrated-leverage route: multiplier analysis on the signup cohort against month-2-retained customers shows that signups who invited a teammate in their first session convert at 6x baseline. The team builds a path that detects this intent earlier (from signup form signals) and surfaces team invitation as a first-session affordance for the matched cohort. Sustainably-paying customers per signup goes up; the cohort-average rate may move less than the first route's 3 points.
The first route moved the metric. The second amplified the pattern that produces sustainably-paying customers.