Crowdfunding Dynamics: Threshold Models and the Bandwagon Backer

Threshold models of collective action

Mark Granovetter's 1978 threshold model is the ancestor of every bandwagon simulation: each individual has a personal threshold — the fraction (or number) of other people already participating that would make them join too. A population with a smooth spread of thresholds can tip from near-total inaction to near-total participation with only a tiny nudge, because each new joiner lowers the bar for the next-most-reluctant person. Crowdfunding maps onto this almost exactly: a backer's threshold is how much of the funding goal they need to see already pledged before they trust the campaign enough to pledge themselves.

Why funding curves are U-shaped

Empirical studies of Kickstarter-style platforms (Kuppuswamy & Bayus is the widely cited one) find that the daily pledge rate on a typical successful campaign is U-shaped: a strong opening burst from the creator's own network and early adopters, a quiet middle stretch, then a second surge as the deadline approaches and the goal comes into sight. The threshold-model reading is direct — the opening burst is the low-threshold backers who need no social proof at all; the trough is everyone else waiting for more proof than currently exists; the closing surge is a mix of proof (visible momentum near the goal) and a hard deadline collapsing everyone's patience at once.

This means a campaign's fate is often decided in the first 24-48 hours: campaigns that clear roughly 20-30% of their goal early enough are statistically far more likely to fully fund, because they cross enough backer thresholds to keep the visible momentum climbing through the trough instead of stalling in it.

All-or-nothing as a coordination device

Most reward-based platforms use an all-or-nothing rule: if the goal is not met by the deadline, no money changes hands and all pledges are refunded. This looks like it only protects backers, but it also solves a coordination problem for the creator — without it, a backer who suspects the project will fall short of a viable budget has an incentive to wait and see, since a half-funded project might still ship something. All-or-nothing removes that ambiguity: a pledge is a bet on the group crossing the line together, or nothing happens at all, which is exactly the condition a threshold model needs to produce a clean tipping point instead of a smear of partial outcomes.

Bandwagon signal vs quality signal

A subtlety worth separating: visible progress toward a goal can mean two different things to a backer, and real campaigns blend both. It can be a bandwagon signal — pure social proof, "other people trust this so I will too," independent of any information about quality. Or it can be a quality signal — early backers who evaluated the project carefully are revealing information the later backer does not have. A model with pure bandwagon dynamics can fund a mediocre project purely on early momentum; a model that mixes in a true quality signal is more realistic and also more fragile, since a single well-informed skeptic sharing bad information can crash a campaign that would otherwise have crossed its threshold cascade.

A minimal simulation loop

The whole dynamic reduces to updating one shared state — percent funded — and letting each backer compare it to their own threshold every tick:

for (const backer of pool) {
  if (backer.pledged) continue;
  const visibleProgress = totalPledged / goal;      // 0..1+
  const p = sensitivity * Math.max(0, visibleProgress - backer.threshold);
  if (Math.random() < p) {
    backer.pledged = true;
    totalPledged += backer.amount;                  // feeds back into next tick
  }
}

Frequently asked questions

Try it live

Everything above runs in your browser — open Crowdfunding Dynamics and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.