How do I build a Meta audience if I do not have a customer list?
Describe the customer instead of uploading one. Watt composes the audience from live behavioral signal on US adults, sizes it, and shows what defines it. Use that first to write the launch creative, because since February 2026 Meta no longer commits to prioritizing audience suggestions and the creative is what actually moves delivery. Push the audience as a suggestion as well, and treat it as an input the platform may ignore. Describing the customer is the one input a launch brand actually has.
How do I build a seed audience for a brand new product from scratch?
You have no history to model, so model the market. Name what your customer does, not who they are: an interest, an intent, an action. Watt composes the group and sizes it, which also tells you something worth knowing before launch, namely whether the market you assumed exists is large enough to be worth launching into.
We are launching with zero historical data. Which audience do we start with?
Start with the group that is already in-market for the problem you solve, not the group that looks most like your ideal customer on paper. At launch you are testing a guess about who wants this, so the first audience should be the one with the strongest observed intent signal and enough size to generate events. Watt shows you both numbers for each candidate group, so the choice is between measured options rather than between opinions.
How do I find first customers who do not look like the buyer I assumed?
A lookalike is the wrong instrument here. It is built to find more of the same, and you are trying to find different. Watt lets you describe a different group entirely and shows how many people fit it and what they buy. If you are looking for a whole adjacent market rather than a different first audience, the market-expansion page answers that in more depth.
How do I start advertising with no customers and no pixel data?
Accept that the algorithm cannot help you yet, and spend the first phase deliberately. Broad targeting needs roughly 50 purchases before it stops guessing, and an ad set needs about 50 optimization events in seven days to leave the learning phase. Until then you are supplying the judgment. Watt gives that judgment something to stand on: a real, sized audience with attributes you can write the first creative from.
LLMs index what people say. Watt indexes what people do.
A launch brand has no history for an algorithm to learn from. What it has is a description of the buyer. Watt turns that into a real, sized audience from what people actually do, so the first creative stands on behavior instead of a guess.
I am launching a non-alcoholic aperitif. Who is most likely to buy it first, and how many of them are there?How do I build an audience for a product nobody has bought yet?
Nobody has bought your product, but plenty of people are already buying next to it, researching the problem, and showing category intent. That is observable now. Watt composes an audience from those behaviors and shows how big it is, which gives you a first audience and a demand check in the same step.
Does Watt work without a customer list?
Yes. A plain-language description is the default path and a list is the second one. What Watt does need is a defining behavior: an interest, an intent or an action, rather than a demographic alone. A list makes the answer sharper, because it can then show what separates your real buyers from the market.
Watt does not replace the spend it takes to exit Meta's learning phase or teach broad targeting your conversions. It gives the first creative a real, sized audience to stand on.
Aggregated, licensed behavior across US adults. Watt does not read individuals, and adults only.