What market observation can and cannot do

Observation produces an overview, not certainty. On the difference between a defensible assessment and a forecast dressed as analysis.

Market observation has a reputation as a sober activity. It becomes one, however, only through a distinction that is frequently omitted in practice: between what was observed and what is concluded from it.

Observation and inference are not the same

That a number of companies invested in a technology over a given period is an observation. That the technology will prevail is an inference. The first can be evidenced, the second cannot.

Where both are presented in the same paragraph and the same tone, the inference borrows the credibility of the observation without having earned it. That is the commonest defect in market reports, and it is usually not deliberate.

Hindsight bias

Market observation rests on what has become visible. What works becomes visible disproportionately. Ventures that were discontinued rarely leave a trace that could be analysed.

A systematic distortion follows: the observable sample looks more successful than the population. A statement about success rates based on visible cases necessarily overstates them.

Temporal resolution

Observation captures states, not movements. Two readings a year apart show a change, but not its course and not whether it continues. From two points, any curve at all can be drawn.

That is not an argument against comparison over time. It is an argument against deriving a direction from a short series and extending it.

What remains

What market observation reliably provides: it shows which questions others have asked, which approaches exist, and where effort is concentrating. That is useful, because it orders your own search.

What it does not provide: a statement about what happens next. Demand that, and you receive a forecast dressed as analysis — and pay the price that the two can no longer be told apart.