A signal proxy is a measurable trace standing in for a quality that remains unseen.
Many important qualities are difficult to observe directly. Trust, influence, satisfaction, attention, and wellbeing cannot be reduced to a single immediate measurement. Institutions therefore use signals that appear to correlate with them: ratings for quality, followers for influence, clicks for interest, or market value for significance.
A proxy is not automatically false. It is a practical approximation whose usefulness depends on the relationship between the visible indicator and the hidden quality. That relationship can be strong in one context and weak in another. It can also decay when people learn how the proxy is measured and begin optimising the number rather than the underlying reality.
The central error is reification. A rating stops being treated as evidence about quality and becomes quality itself. A high market value becomes proof of importance. A click becomes proof of interest. Once the proxy takes the place of the thing it was meant to indicate, measurement begins to organise behaviour rather than merely describe it.
This is why signal proxies need interpretation, not worship. A number can compress complexity and make comparison possible while still hiding context, distribution, incentives, and what was never measured. Credibility and market capitalisation are familiar examples of qualities often inferred through signals whose validity depends on the question being asked.
Source: Scholarly foundation: Kenneth A. Bollen, Structural Equations with Latent Variables (1989), on observable indicators and latent constructs; supplemented by a public synthesis of proxy signals and metric reification.