I stayed up too late on Tuesday, like most people I know, and by Wednesday morning there were four newsletters in my inbox explaining what the polling got wrong. All of them confident. None of them agreeing with each other.
I am not going to write about the politics. I want to write about the instrument. The best funded, most methodologically obsessive measurement operation in this country spent two years asking people what they intended to do, and the answers did not reliably predict the doing. Again.
Which should unsettle anyone who has ever put a segmentation deck in front of a CMO.
If professional researchers with enormous stakes and decades of method cannot dependably turn stated intent into behavior, your twelve hundred respondent brand tracker is not evidence, it is a mood ring.
Consider everything that has to go right for a survey to mean something. First, the people who answer have to resemble the people who do not, and they do not, increasingly, because response rates have collapsed and the willing respondent is a specific kind of person. Second, those people have to describe themselves accurately, which nobody can do, because we all narrate ourselves as slightly better and slightly more deliberate than we are. Third, you weight the sample to correct for the first two problems, and the weights are a theory, and the theory is usually the part that breaks.
Brand research carries all three failure modes and none of the accountability. An election gets a scoreboard on a Tuesday in November. Your persona deck never gets graded. It gets laminated.
So the response is not to stop researching. It is to stop treating stated preference as fact and start weighting revealed behavior much harder. What did people actually buy, return, reorder, abandon in a cart, recommend in a group chat without being asked. Salt & Straw does not run a survey to pick flavors, it runs limited batches and watches what sells out by Thursday. That is a survey with money on the table.
Second, hold your findings loosely enough to notice when they expire. The most dangerous research in any organization is not the flawed study. It is the perfectly good study from four years ago that nobody has reopened.
Third, go be in the room. Watch ten customers use the thing. It is not statistically valid and it will still tell you more than the tracker.
Your persona deck never gets graded. It gets laminated.
It rained all Wednesday, which the forecast did get right. I bought a book at Powell's I do not need and went home and sat on the floor. Truffles has extremely strong preferences and states exactly none of them. She simply goes and sits on the thing she wants. I have come to think of her as the most reliable panel I have access to.