The Bass Model: How Innovations Spread Through a Population

Two kinds of adopters

In 1969 the marketing researcher Frank Bass proposed that every potential buyer of a new product falls, in aggregate, into one of two behavioral camps. Innovators decide to adopt independently of what anyone around them is doing — a magazine ad, a press release, or simple curiosity is enough. Imitators adopt because they have seen the product in the hands of people who already own it; their decision is driven by social contact, not advertising. Almost no real buyer is purely one or the other, but a market-level average of the two turns out to fit adoption curves for hundreds of products astonishingly well, from color televisions to electric vehicles.

The differential equation

Let M be the eventual market size, N(t) the number of people who have adopted by time t, p the coefficient of innovation and q the coefficient of imitation. The rate of new adoptions is:

dN/dt = ( p + q * N/M ) * ( M - N )

p        = external influence rate (advertising, media, curiosity)
q * N/M  = internal influence rate, proportional to the fraction
           of the market that has already adopted
(M - N)  = the remaining, not-yet-adopted pool

Read it left to right: the pool of people who have not yet bought shrinks over time, and the rate at which they convert is a mix of a constant trickle (p) and a contagion term that grows as more people already own the product (q·N/M). The equation has a closed-form solution, a logistic-like curve skewed by the ratio of p to q, and it can be fit to early sales data with nonlinear least squares or maximum likelihood.

Finding the peak: t*

The most useful single number the model produces is the time of peak new adoptions — the point where dN/dt stops accelerating and starts decelerating, even though cumulative adoption keeps climbing toward M. It has a closed form:

t* = ln(q / p) / (p + q)

At t = t*, N(t*)/M = p / (p + q)

Because q is usually far larger than p for products that spread by word of mouth, t* arrives once a fairly small fraction of the market — often well under half — has already adopted; the imitation term takes over and the rest of the curve is steep. Manufacturers use t* to time capacity expansion: build up production ahead of the peak, not after it.

Fitting real markets

Bass's original 1969 paper fit the model to eleven durable goods including room air conditioners, television sets and clothes dryers using only a handful of years of early sales, and the model has since been re-estimated for products ranging from VCRs to smartphones to plug-in electric vehicles. Typical estimated values cluster around p ≈ 0.01–0.03 and q ≈ 0.3–0.5: word of mouth usually dominates external influence by an order of magnitude once a product has any social visibility at all. Extensions add a third parameter for repeat purchase, split p into separate advertising and non-advertising components, or let M itself grow over time as incomes rise or a category expands (smartphones absorbing what used to be separate camera and music-player markets, for instance).

Where the model breaks

The Bass model treats the market as one homogeneous pool with a fixed ceiling M, no competition, and a single once-only purchase decision. Real markets have competing products splitting the same q, price cuts that pull the curve forward, supply constraints that cap early sales below what demand alone would produce, and — in categories like smartphones — an M that keeps expanding as the product itself changes. None of that invalidates the core insight: even a two-parameter model of "some people decide alone, most people copy their neighbors" reproduces the S-shape seen again and again across a century of consumer technology.

Frequently asked questions

Try it live

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