Composite Indices and Why They Mislead
Combining several measures into one index is tempting for reporting and destroys the information that made them useful.
A productivity index combines output, quality, utilisation and whatever else into a single figure. It is popular with reporting and almost always a mistake.
The practical lesson in “Composite Indices and Why They Mislead” is to connect every number to a decision and retain the context behind it. Teams exploring workforce analytics software can review workforce analytics tools for management reporting as one source of operational evidence, provided the purpose is disclosed and the interpretation is tested with the people affected.
What is lost
Which component moved.
For an independent perspective related to “Composite Indices and Why They Mislead”, consult the Harvard Business Review productivity collection; it offers a useful external check on definitions, governance and the assumptions built into a proposed measure.
Whether two components moved in opposite directions and cancelled.
The units, which become meaningless after normalisation.
And the ability to act, because an index has no action attached to it.
The cancellation problem
Output up, quality down, index flat.
Reported as stable, which is the opposite of what happened.
This is not an edge case — it is the predictable result of combining measures that pull against each other, which the counterweight note deliberately sets up.
The weighting problem
Somebody chose how much each component counts.
That choice is a judgement about what matters, embedded in arithmetic and invisible afterwards.
Change the weights and the index changes with no change in the work, which makes comparison over time dependent on nobody having adjusted them.
Why it survives anyway
One number fits on a slide.
It allows ranking, which somebody wants.
And it looks rigorous, which the ratio note explains is the general hazard of arithmetic applied to approximations.
What to do instead
Report three or four measures, unweighted, on one page.
Volume, quality, flow, and one about cost if the question is budgetary.
A reader can hold four numbers, and each of them has an action attached.
If an index is unavoidable
Publish the components and the weights beside it.
Never report the index alone.
Record the weighting and the date, so a change in composition is distinguishable from a change in reality.
And never rank teams by it, which the comparison note covers.
The variant worth watching
Vendor-supplied productivity scores do exactly this with weights you cannot see.
Which is the same objection, with the additional problem that the judgement about what matters belongs to a supplier.
Its own adjacent subject, and the principle transfers directly.
The honest sentence
"These four measures moved like this, and here is what we think it means."
Longer than an index and the only version that supports a decision.
What to check
Does your reporting contain a composite index?
Can you see its weights?
Has anything been reported as stable while components moved opposite ways?
And could you replace it with four numbers on one page?