Brand health figures presented in marketing reviews come from a continuous survey with a specific structure. The construction of that survey determines what the numbers mean.
Sampling defines the population being measured
A tracker interviews a sample intended to represent a defined group, usually category buyers in a market rather than the general adult population.
Getting that definition wrong changes everything downstream, since awareness among all Americans and awareness among people who buy the category are different quantities.
Samples are then weighted to match known population characteristics, which corrects composition but cannot repair a poorly chosen definition.
Question order creates the funnel
The questionnaire moves from unprompted to prompted deliberately, asking which brands come to mind before showing any list.
Unprompted recall is the hardest measure and the most meaningful, because it reflects retrieval without help, which is closer to how purchase decisions begin.
Prompted awareness follows, then consideration, then usage, and each step must come after the previous one because seeing a list contaminates any unaided question asked later.
Continuous measurement beats periodic waves
Interviewing a small number of people every week produces a smooth series in which advertising activity can be related to movement over time.
Large surveys run twice a year produce two points, and any change between them could be caused by campaigns, seasonality, competitors or sampling variation.
Continuous designs cost more to operate but are the only version that supports comparing periods with and without media activity.
Movement is slower than campaign cycles
Brand measures change gradually because they reflect accumulated memory across years of exposure rather than reaction to the most recent flight.
Weekly figures also carry sampling noise, so a small movement in one week is frequently indistinguishable from random variation.
Reading rolling averages over longer windows separates genuine trend from noise, and prevents campaigns being judged on a fluctuation.
Comparability constrains what can be changed
The value of a tracker comes from consistency, since the series only supports comparison if the questions and method stayed the same throughout.
Changing wording, adding a brand to the prompted list or switching sample provider can shift levels enough to break the history.
Serious changes are therefore run in parallel with the old design for a period, so the size of the discontinuity is known rather than guessed at afterward.