Build the two hundred row prospect list without overspending

How do you stop a search costing more than it returns?

By making the first one small.

Search takes a size parameter that defaults to ten and caps at a hundred, and every match is billed as People Data Labs credits on your account. So the sequence that works is a ten-row probe, a look at whether those ten are the people you meant, a correction to the query, and only then the full pull with paging.

A companion suits that loop because it reads the ten rows instead of counting them. It spots the consultants, names the filter that removes them, and comes back with the corrected query before you have spent anything else. That conversation costs a hundred matches instead of the two thousand a slow discovery costs.

Why write the query as SQL?

Because you can read it at the approval step before it is billed.

The person search accepts a SQL string covering job title role, company size and country, or an Elasticsearch query object when SQL cannot express what you need, and a nested boolean object is not something anybody checks properly at 4pm.

Enrichment is the cheaper half and comes first whenever you already hold a key. An email, phone number or profile URL returns work history, education, location and socials. A domain, name or ticker returns industry, size, revenue, founding date, headquarters and technologies.

Get the enriched rows into your CRM

Two hundred enriched rows are worth something when they land in the CRM with the reasoning attached. A companion checks the rows that look wrong against the live web and writes them into the systems you are signed into, with record-creating writes waiting on your approval.

That is sales prospecting with the enrichment step already finished, and the same shape serves recruiting when the search is for candidates. The ten-row probe comes to you first, so a filter that is subtly wrong gets fixed at ten rows rather than at two hundred.

Ask whether a row is already a customer

Whether somebody is already a customer, or whether the account was disqualified in March.

Those are questions about your systems, and a companion answers them by reading your CRM in the same session as the enrichment, which is what a script against the API cannot reach.

The dataset side is exactly four calls: person search, company search, person enrich, company enrich. There is no bulk endpoint, no IP lookup, no filter autocomplete, no taxonomy tool and no streaming, and nothing writes back to People Data Labs.

• Search people or companies with SQL or an ES-style query; ten by default, a hundred maximum per call.
• Enrich a person from an email, phone, profile URL, or a name plus company.
• Enrich a company from a website, a name or a ticker symbol.
• No bulk, IP, autocomplete or taxonomy endpoints, and no writes of any kind.

Frequently asked questions