The traditional definition of an output is that it is a ‘good or service’ produced by an intervention (a project, program or organization). In some instances it is defined as a ‘final’ good or service to distinguish it from internal processes.
Within outcomes theory an output is formally defined as a box within an outcomes model which has three features: 1) it is measurable; 2) it is controllable by another box within the outcomes model; and, 3) it can be viewed as a good or service.
One of the most useful things about an output within outcomes work is that because it is measurable and controllable it does not have an ‘attribution’ problem. An attribution problem is when merely measuring that a box within an outcomes model has happened does not prove who made it happen. For controllable boxes such as outputs, their mere measurement means that we know who made them happen.
To understand this definition more clearly, it is useful to understand the different features of boxes that can appear within outcomes models.
In some traditional ways of drawing outcomes models (in particular the type of outcomes models called logic models) outputs are restricted to a particular column in the model and in some instances there is only allowed to be one column of outputs and no other boxes prior to a column of outcome boxes.
This can artificially constrain the modeling of what is actually happening in an intervention. In some interventions this may make sense where an output is very close to its outcome on the right-hand side of the outcomes model. An example of this is an vaccine program where the output - giving an injection is very close to the outcome of reduced sickness and deaths.
However in other interventions, for instance health promotion interventions aimed at changing legislation regarding health issues (e.g. tobacco, alcohol, sugar) there may be many steps between an output such as producing a Youtube advocacy video as part of mobilizing a community and the ultimate legislative change. This is shown in the example below.
The traditional approach which limits the number of intermediate boxes allowed between outputs and outcomes boxes can make it difficult to portray the complexity in the second type of situation described above. More information is available comparing the pros and cons of traditional logic models with multi-layered outcomes models.
A Three MInute Outcomes video is available on the difference between an output and an outcome.
[This is a provisional definition of an output within outcomes theory. If you have any comments on how it could be improved please send them to us via our comments page}.