Every data provider is the cheapest for something and the most expensive for something else. Running them in the wrong order on a large list is how enrichment budgets disappear without anyone noticing.
The order is the product
Sourcing the same list from a single provider and sourcing it through a cascade produce roughly the same rows at very different prices. The difference is entirely in what gets asked of whom, and in how much of the list never needs to be bought at all because you already hold it.
So the skill treats provider order as a decision to be made per run rather than a setting configured once. A list of European mid-market operations leaders and a list of US enterprise engineering managers do not have the same cheapest path.
The cost gate
Nothing is spent before you see a number. The preview stage returns match counts per provider without pulling rows, which costs nothing, and the estimate is presented for approval. This exists because the alternative, finding out what a build cost after it ran, is how teams end up rationing the thing that generates pipeline.
Two shapes of build
People direct, where you know the titles you want and go straight at them. And company first, where accounts are discovered against firmographics and signals, then decision makers are found inside each one. The second is what an account based motion needs, and it is a genuinely different query, not a filter on the first.
Discovery only
It builds the list and stops. Nothing is contacted, and the list does not enter a campaign until a separate signed-off step puts it there.
Runs well with
Skills compound. These are the ones we usually install alongside it.