Growth Modeling

Good piece by Vincent Granville on growth modeling in DSC.  Non technical.  Intro below with details at the link.

Growth Modeling for Business Managers and Executives
Posted by Vincent Granville  in DSC

You don't need a sophisticated model nor advanced machine learning techniques to quickly get a high level picture and trends for bottom-line business metrics. Not only the concepts explained here are easy to grasp, but while being high level, it nevertheless includes granular effects. The methodology presented here was used in business contexts in the past, when I was working with enterprise executives, particularly finance people, to assess the overall health of their business, and the short and long term impacts of new initiatives to boost growth. .

The model is available as an Excel spreadsheet, driven by four main parameters, as illustrated below. The growth can be in revenue, users, or any other fundamental metric. Time periods are measured in days when assessing the impact of an advertising campaign, or in months when assessing revenue growth caused by a new initiative. It typically involves the following dynamic:

New growth occurs at each time period, for instance new users.

It accumulates over time: new users become regular users, some of them eventually disappear -- this can be factored in in the growth curve.

There is usually a time lag between an action and a reaction:  the effect of TV advertising campaigns may peak after a while (not immediately) and eventually decay.   .... "