From a list
Bring a list you already own and Watt expands it, learns what defines it, or slices it by intent — and every play chains with the rest of the build.
You don't have to start from an idea. Bring a list you already own — customers, leads, accounts, event registrants — and say what you want to do with it:
/watt:audience expand my list to every match/watt:audience find more people like my best customers/watt:audience rank my list by who's most in-market for solar/watt:audience which of my customers are also weekend hikers/watt:audience take my purchasers who live in Texas, and expand to their households
What you can bring
Paste identifiers straight into the chat or upload a CSV. Watt matches on common identifiers like emails, phones, names, and addresses — any mix — and you can ask /watt:help for the full list of identifiers at any point. For a CSV you choose which columns to match on; the rest ride along untouched. A few practical notes:
- More identifiers, better matches. Watt supports a wide range of primary and secondary matching criteria, and the strongest way to match is to bring several per person — on a multi-identifier match Watt blends them into a single quality score it hands back.
- Counts stay honest at every step. You always see "8,400 identifiers in, 7,900 people matched" — never a silent shortfall.
- Your list stays yours. Matching never shows you graph records for your people — you get the matched set and its counts; individual data only exists downstream, in an export you explicitly confirm. See Data handling.
What a list is good for
A list can be the seed a lookalike grows from, the set a ranking sorts, or the thing you want to understand. Every play starts the same way — Watt resolves the list against the graph and hands back the honest match count — and then the play does its work. Three plays cover the ground, and they chain — with each other, and with everything a description-driven build can do: take the people who bought from you, keep the ones who live in Texas, expand to their full households, and advertise to that.
Expand to every match
The widest set: every person your identifiers plausibly point to, with a match confidence on every row so the long tail is never hidden. Use it when addressable reach is the goal, or when a strict match leaves too much of your list behind and you'd rather lift the match rate and pick the confidence cutoff yourself.
Households fall out of this. An address matches everyone living at it — so expanding a list that carries addresses brings back the co-residents, the whole household, not just the named person.
Learn what defines it
Don't work the list itself — learn from it. Watt reads your list for the signals that set these people apart from everyone else: the durable identity (interests, demographics, life stage) and what they're currently in-market for. You get that back as a signal stack you tune — keep what generalizes, drop what just means "they're already my customer" — then build a fresh audience of more people like them through a normal description-driven build. The list is the seed; the audience it grows is new people.
Slice and dice it
Keep the people on the list, but work the set. The core move is scoring: you describe what to score by — "in-market for solar", "showing home-buying intent" — and Watt scores every matched person by how many of those signals they express, returning the whole list ranked highest to lowest. Lead-scoring without leaving the chat. From there, shape it:
- Overlay it against signals you already hold — a stack from a previous build, a pool from exploring the graph, the signals learned from another list — the score doesn't need a fresh description, any signal set works.
- Cut it to the slice that matters — "only the ones showing real intent" — when you want the shortlist, not the ranking.
- Intersect it with anything you can describe — "my purchasers who live in Texas", "my customers who are also weekend hikers" — the same signals a description-driven build runs on, applied to your people.
- Group it along a dimension you name — "break my customers down by region", "split them by life stage" — the set falls into its best-concentrated cells, each one workable on its own.
Whatever the play lands, the set follows the shared flow — analyze who they are, or activate them as a platform-ready file.