GTMGTM

Find · Sourcing

Two sources, because there are two questions

A prospect corpus answers who matches your ICP. Your own LinkedIn account answers who just did something interesting. Neither replaces the other, and they run on separate clocks so one failing vendor never takes the other's leads with it.

Contacts14,208
Contacts1,284Series-A fintech opsImport CSV
NameCompanyStatusIntent
PRPriya RaghavanNorthwindQueued5
DODaniel OkoyeCadenceQueued4
MLMara LindqvistFernwehQueued4
ABAisha BelloThreadlineBooked5
THTomás HerreraHalyardNew3
EMElena MarchettiSablefishNew2
JWJonas WeberKestrelNew1
RCRavi ChandranLoomisNew1
1–8 of 1,284Page 1 of 161

What is actually going on

Every claim below is followed by the mechanism under it, because a claim without a mechanism is a slide.

01

The corpus pull is sized before it runs

Under twenty-five matches is a typo, not a small market. Over four million is not an ICP, and paging into it would import whoever the database happened to sort first. The count is checked before a single credit is spent.

02

Filter values are resolved by the vendor, never guessed by a model

The United States is id 4000012 in one corpus and 1 in another. Asking a model to guess would buy a silent failure mode: a wrong id matches nothing and reports success. The ICP is stored as human phrases and resolved through the vendor's own lookup, with the match count attached.

03

LinkedIn search picks one of four instruments

Classic LinkedIn has no single rich people search, so the signal kind chooses the tool: post engagement for a keyword, authors for a news mention, job postings for a hiring trigger, people search for everything else. Engagement beats authorship. A post has one author and often hundreds of engagers.

04

Job postings are the only route to a company domain

Classic job search is the one search that returns a numeric company id, and the company profile is the one place a real website exists. Hiring leads therefore arrive with a company name, a domain and a headquarters country already attached, instead of inferred.

Step by step

How lead sourcing runs

  1. 1

    Resolve

    Your ICP phrases become the vendor's own filter ids, with a live match count.

  2. 2

    Size

    Too small is a typo; too large is not an ICP. The walk is refused rather than run badly.

  3. 3

    Walk

    Up to 200 candidates per workspace per daily tick, with a cursor that moves forward across ticks and resets when the ICP changes.

  4. 4

    Write

    Every candidate is revealed before the insert. Only the people the corpus answered for with an address become contacts.

The numbers

Facts, not adjectives

Every figure here is a real value from the running product. A cadence from the scheduler, a price from the credit table, a limit from the ramp. A marketing page that invents a limit is a support ticket with a delay on it.

FindSourcing
Corpus size
700M+ contacts, licensed
Corpus pull
Daily, 200 candidates per workspace
LinkedIn sweep
Every minute, per active signal
Candidates per signal
Up to ~120
Search cap
1,000 results. LinkedIn's own ceiling
Provenance recorded
source_system, collection_method, collected_at

Lead sourcing, honestly

Do I need a Sales Navigator seat?

No, and nothing in the product assumes you have one. Classic LinkedIn's four searches plus the licensed corpus are the whole sourcing surface.

What happens if I have not connected LinkedIn?

Signals go idle and say so. Onboarding is not blocked on it, and corpus sourcing keeps running. LinkedIn is the slowest channel to connect and the most likely to fail, so nothing waits on it.

Lead sourcing is one part of it. The rest runs too.

Sourcing, enrichment, intent, copy, sending, replies and reporting are one product with one bill and one gate. Connect a channel and the whole motion starts.

Start with sales. $2,000 or $5,000 onboarding, then at least $250 monthly credit funding.