voltage-effect
Voltage Effect
Overview
The voltage effect is John A. List's name for what happens to most promising ideas when they scale: they lose voltage — they fail, shrink, or reverse — because the conditions that produced the small-scale win do not survive being scaled. The discipline is to predict scalability before you scale, by interrogating an early win against five vital signs: (1) false positives — the pilot result was never real (underpowered, p-hacked, or a fluke); (2) an unrepresentative population — it worked on early adopters or a hand-picked group that will not generalize; (3) an unrepresentative situation — it worked in one context, channel, or market that will not replicate; (4) spillovers — general-equilibrium effects (congestion, saturation, competitive response) that only appear at scale; (5) the supply-side cost trap — unit economics, talent, or operations that break as volume grows, so the idea is unscalable even when demand is real.
List frames the burden of proof plainly:
"The vast majority of ideas—no matter how great they seem at first blush—don't scale. And it's often not until we take these ideas to scale that we discover their fatal flaws." — John A. List, The Voltage Effect (2022)
List's larger point: what makes an idea scalable is not the same as what makes it seem promising in the first place.
The correct default is skeptical: an idea is presumed non-scalable until it survives all five vital signs. List's scalable levers are the other half: high-fidelity implementation (the scaled version must deliver what the pilot delivered, not a diluted copy), marginal thinking (decide on marginal cost and marginal benefit at scale, not on pilot averages), and designing for scale from day one (build the population, the situation, and the cost curve you will need at volume into the first test).
Compose with neighbors. Run voltage-effect before economies-of-scale: economies-of-scale asks whether cost per unit falls with volume, but that is only vital sign #5 — clearing it while ignoring #1–#4 scales a fluke efficiently. Use it instead of raw optimism about network-effects: claimed network effects are a scale-only benefit that must be demonstrated, and their mirror image (congestion, saturation) is exactly the spillover of vital sign #4. Use it after pmf-crossing-the-chasm: the chasm names why early-adopter traction fails to generalize; voltage-effect vital sign #2 makes that failure a checkable gate. It shares machinery with representativeness-heuristic (a vivid pilot is a prototype that overrides base rates) and with lean-startup (validated learning is the antidote to false positives — but voltage-effect insists the learning generalize, not merely replicate the same test).