Better late than never.

[ First recorded about 1200 ]

Microeconomics and marketing literature features three basic models of adoption and purchase of a product on the market:

- Fourt–Woodlock
trial and repeat model

The The model is composed of a trial and a repeat parts. The trial target segment is usually supposed to decay exponentially. - Parfitt-Collins
steady-state model

The model is an alternative formulation of the repeating purchase by taking in account returning customers. - Bass
diffusion model

The adoption rate is positively influenced by the cumulative number of adopters. Therefore, the model is sometimes called epidemic.

The three models have been united into a single model. With some extensions, the TRD - Trial and Repeat Dissipation model was developed.

- The original models have been reformulated as stochastic by changing the rate parameters into probabilities per time unit.
- Length of the repeat period has been added as a parameter. The length has been set equal to the mean time of repeated purchases with exponential dissipation (decay) of purchase probability. A user of a product becomes a trier when the period of repeated purchase ends. The ratio of returning users can be set as the switchback ratio.
- The model does not have an analytical solution. The numerical Runge-Kutta integration procedure is implemented.
- A sales timeline can be composed of at most 105 time periods of arbitrary lengths. Each time period can have different parameters treated as constant in the time period.
- A simple tool allowing to estimate projected awareness at some advertising intensity relative to some known intensity, has been added. (The method comes from Arthur D. Little.)

The MS Excel-97 simulator was finalized in 2005. It appeared very soon that the original intent to use probabilities obtained from conjoint measurements as parameters was unrealistic. Quite generally, conjoint derived parameters are essentially relative ones, but their absolute values differ a lot from those on the actual market. Other required parameters, namely mean time of the repeat cycle and switchback ratio, are hardly obtainable from a questionnaire-based research. A consumer panel might be of some help, but the best estimates can be obtained from detailed market data. Such data are hardly ever revealed.

The simulator might be helpful to product managers, students of marketing or microeconomics, and to all who are interested in studying influence of changing market conditions on sales.

Below is a simulator produced graphics of a textbook-like example. A little known product is subject to a well designed promotion campaign which is leading to increase of sales.

The simulator, a description of its functionality and several examples are available as download.

The VBA part of the simulator is protected by a
password. Please note the simulator is a historic product that has
never been or asked to be used commercially. The documentation on the
implementation is unavailable.