How it works
Deriverso computes a single value for each topic by combining its measures. The rules are simple, but they change depending on how your topic is configured.
A topic's computed value depends on its measures having values. Each measure's value can be set manually or pulled automatically from an integration. Until at least one measure has a value, the computed value shows "—".
With an equation, all variables must have at least one assigned measure with a value before the equation can be evaluated.
When a topic has no equation, all of its measures are combined into a single weighted average:
Each measure carries a weight between 0 and 1 that reflects your confidence in its relevance. A higher weight means that measure pulls the result more strongly toward its value.
Weights are relative, not absolute. They control how measures are mixed together but never amplify or dampen the final value beyond the range of the inputs. A single measure always equals its raw value regardless of its weight.
When a topic uses an equation, the computation changes. Measures are grouped by the equation variable they're assigned to, and each group is reduced to a single value before the equation is evaluated.
Each measure is assigned to one of the equation's variables.
If a variable has multiple measures, their weighted average becomes that variable's value. If a variable has only one measure, its raw value is used directly.
The resolved variable values are plugged into the equation expression to produce the final topic value.
Equation: score = (A + B) / 2
Weights don't always change the result. Here are the three rules:
Weights are always applied. All measures are peers, so weights determine how much each one influences the shared result.
Weights activeWeights are applied within the variable group. Measures sharing a variable only influence each other — weights don't change the variable's impact on the equation.
Weights activeThe measure's raw value is used directly. Reducing a value because you have less confidence in it would distort the equation — the measure is the variable's only source, so its raw value is used. You can still set a weight — it will take effect if another measure is added to the same variable.
Weights inactive