Not everybody
A business wastes money the moment it tries to reach everybody, and the waste is invisible because the reports still fill up.
The reason segmentation gets done badly is that the easy variables are the useless ones. Age, location, company size and industry are available in every ad platform, which is why every plan uses them, and there is a large and specific piece of evidence about what they can carry. In that study of 52 product categories, demographic variables improved the model for whether someone used the category in 44 of them, and for which brand they preferred, in none of them. Not a weak association in a few categories. None, in 52.
That is not an argument against demographics. It is an argument for knowing what job they do. They are a filter on who is plausibly in the market, which is exactly how a media buyer should use them. They are not an explanation of behaviour, and every plan that builds a persona out of them and then writes messaging from it has confused the two.
What does carry the difference is the set of things a demographic table cannot hold.
- What triggered the search. Nearly every purchase has an event behind it.
- What they are using now, and what it costs them.
- What they are afraid of. Usually a specific past failure rather than a general risk.
- What they would have to explain to someone else in order to buy.
- Where they went to find out, and what they typed.
- What would make them stop halfway.
A buyer account that contains a need, a motivation and a helpful demeanour has been written by the marketer. Real buyers carry a hesitation: a price they cannot justify, a person who has to approve it, a previous product that burned them, a suspicion that they will not use it. If you cannot state your buyer's hesitation in their own words, you have not spoken to enough of them, and every piece of content you write will be aimed at a version of them who was always going to buy.
Personas deserve a specific warning, because they are the standard output of this exercise and the evidence for them is thin to the point of absence. The academic survey of the persona literature describes the field as taking a position of advocacy and lacking objectivity, notes that the accuracy of personas is difficult to validate, and concludes that the claimed benefits remain potential. No study was found measuring revenue, conversion or campaign outcomes from persona use, in either direction.
That does not make them useless. It makes them an internal alignment device with an unmeasured effect, which is a fine thing to be as long as everyone knows it. The failure mode is a persona with a stock photograph and a name and a favourite coffee shop, treated as a research finding because it is laid out like one.
This course asks for buyer accounts instead, and the difference is not cosmetic. An account is tied to specific purchases that happened, it names the people it was built from, and it carries the hesitation. Three accounts for one offer, each with the trigger, the alternative, the hesitation and the reason they went ahead. Where two accounts have the same trigger and the same alternative, they are one account and you have two customers, not two segments.
Your ad platform reports that your best-performing audience is women aged 35 to 44 in three cities. What should you conclude?
You have written three buyer accounts. Two of them have different job titles and different industries, but the same trigger and the same current alternative. What follows?
Notes are kept with your account, alongside your progress and your gate claims. The lesson itself is readable without one.