# Map data referral opportunities in your network | SourceX

Canonical: https://refer.sourcex.si/guides/business-network-data-referral-opportunities
Last updated: 2026-10-27

To map potentially eligible companies in a network, list the companies you already know, then screen each on four facts: at least 50 full-time employees at peak, a qualifying location, years of its own operational records in a named system, and a decision-maker you can reach. Sort them into three buckets: ask the owner, find facts first, or not a fit now. The output is a short list of conversations, not a forecast of income. An afternoon with twenty companies is a reasonable first pass.

## How can a partner map potentially eligible companies in an existing network?

Start from relationships, not databases. A company you already advise, audit, insure or supply is one where you can ask the owner a direct question. Screen on what you know, mark what you do not, and only then decide whom to call.

The live guides for [Salesforce consultants](/guides/how-crm-and-revops-consultants-spot-referral-opportunities-during-salesforce-projects), [HubSpot implementation partners](/guides/how-crm-implementation-partners-spot-referral-opportunities-during-hubspot-projects) and [Intercom consultants](/guides/how-customer-operations-consultants-spot-referral-opportunities-during-intercom-projects) cover spotting signals inside a single project. This page covers the step before: going through everyone you know, once, with the same four tests.

## Step 1: pull the list you already have

Two common starting points:

- **LinkedIn connections.** LinkedIn's Help Center describes requesting a copy of your connections from Settings. Email addresses appear only for connections who allow it, and only 1st-degree connections can be exported. Checked 2026-10-18. A step-by-step version is in [export your LinkedIn connections](/guides/export-linkedin-connections).
- **Your own CRM or client list.** If you use HubSpot, its default company properties include Industry and Number of employees. HubSpot describes Number of employees as the total number of people who work for the company, which is a current count. The program test is 50 or more full-time employees at peak, contractors excluded, so treat the field as a hint, not the answer. To find size signals on LinkedIn itself, see [searching companies by employee size](/guides/linkedin-search-companies-by-employee-size).

Collapse the list to companies, not people, and keep only those where you know someone well enough to ask a direct question.

## Step 2: screen on four tests

| Test | Pass | Unknown | Fail |
|---|---|---|---|
| Size | 50+ full-time employees at peak, contractors excluded | Headcount not known, or only a current figure | Clearly under 50 at peak |
| Geography | In a qualifying country, with English records where required | Records language not known | Outside coverage, or records mostly not in English where English is required |
| Own records | Years of the company's own operational records in a named system | System or years not known | Records belong to clients, or barely exist |
| Decision-maker | You can reach someone who can authorize a licensing discussion | You know staff, not leadership | No route to leadership |

Geography, exactly as the program states it: Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 90% of the data is in English.

Bucket rule: four passes go to **Ask owner**; any unknown without a fail goes to **Find facts**; any fail goes to **Not now**.

## The 20-company screen (Illustrative)

A fictional fractional CFO works through her client and former-client list. Every company and fact is invented. No income is estimated for any row, because nothing about payout is known at this stage.

| # | Company | Size | Geography | Own records | Decision-maker | Bucket |
|---|---|---|---|---|---|---|
| 1 | 220-person freight broker | Pass | U.S. | Pass | Pass | Ask owner |
| 2 | 140-person dental lab group | Pass | U.S. | Pass | Pass | Ask owner |
| 3 | 90-person commercial roofer | Pass | Canada | Pass | Pass | Ask owner |
| 4 | 300-person staffing firm | Pass | UK, English records | Pass | Pass | Ask owner |
| 5 | 75-person engineering consultancy | Pass | Ireland (EU) | Pass | Pass | Ask owner |
| 6 | Regional property manager | Unknown | U.S. | Pass | Pass | Find facts |
| 7 | 60-person IT services firm | Pass | U.S. | Unknown | Pass | Find facts |
| 8 | Specialty food distributor | Unknown | U.S. | Unknown | Pass | Find facts |
| 9 | 180-person contact center | Pass | India, English records | Unknown | Pass | Find facts |
| 10 | 110-person insurance agency | Pass | U.S. | Pass | Unknown | Find facts |
| 11 | 95-person logistics firm | Pass | UAE, records language unknown | Pass | Pass | Find facts |
| 12 | Home health agency | Unknown | U.S. | Pass | Unknown | Find facts |
| 13 | 130-person accounting firm | Pass | U.S. | Unknown | Pass | Find facts |
| 14 | 25-person design studio | Fail | U.S. | Pass | Pass | Not now |
| 15 | 40-person bookkeeping practice | Fail | U.S. | Pass | Pass | Not now |
| 16 | 200-person manufacturer | Pass | Mexico, records mostly Spanish | Pass | Pass | Not now |
| 17 | 150-person outsourced support firm | Pass | U.S. | Fail: client-owned records | Pass | Not now |
| 18 | 70-person startup, founded 2025 | Pass | U.S. | Fail: little history | Pass | Not now |
| 19 | 500-person retailer | Pass | U.S. | Pass | Fail: no route to leadership | Not now |
| 20 | 30-person law office | Fail | U.S. | Pass | Pass | Not now |

Result: 5 to ask the owner, 8 needing facts first, 7 not a fit now. The useful output is the action list: five conversations to request and eight sets of facts to find. Rows in "Not now" stay on the list, because headcount, records or leadership can change.

## Step 3: what to say, and what not to collect

For the five "Ask owner" companies, the first conversation stays at metadata level: which system, what kind of record, how many years, roughly how many records, who owns the system and who could authorize a discussion. The [metadata-only opportunity map](/guides/metadata-only-referral-opportunity-map) gives the exact columns. The [company fit checker](/tools/company-fit-checker) is a quick second opinion. Never request samples or exports, and never introduce a company without the decision-maker's knowledge.

## Limits and unanswered questions

- **No rates are assumed.** SourceX has not published what share of screened companies qualify or convert, and this page does not estimate it.
- **Buckets are your judgment.** SourceX reviews each introduction against current dataset demand, rights and record depth; a company in "Ask owner" can still be declined.
- **Agreement first.** Partners sign the Channel Partner Agreement before submitting introductions.
- **Payments stay apart.** Any compensation for licensed data is agreed between SourceX and the company. A partner's reward is paid separately by SourceX and never deducted from the company's payout.

## Next step

> **Refer a company. Earn up to $100,000.** Default reward: 25% of Actual Supplier Payout, capped at $100,000 cumulatively per referred company. Eligibility and payment conditions apply, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

Once an owner in your "Ask owner" bucket agrees to a conversation, make that one introduction: [register](/register), confirm [who qualifies](/who-qualifies), and read the [terms](/terms). More partner playbooks are in [guides](/guides).

## What should a business network owner know about AI data business network opportunities?

The companies that matter here are rarely AI companies. They are established businesses you already know as clients, former colleagues' employers, vendors and owner-led firms, with years of their own operating records. Knowing someone there is not yet an opportunity. Three checks come first: permission (the relationship holder agrees to ask and the CEO wants the conversation), ownership (the company owns its records) and mandate (your own rules allow it). Map them company by company, then open three permissioned conversations.

Keep three kinds of statement apart when you share the map:

- **Known (program facts, checked 2026-10-26).** Introduced companies need 50 or more full-time employees at peak, original documentation they created, rights to license it and an authorized sponsor. Criteria follow each specific program, so a company that misses one may fit another later. Operating, acquired and wound-down companies can all qualify.
- **Conditional.** Acceptance, license scope and any reward, settled only by SourceX's review and signed agreements.
- **Our thesis, not measured.** The deepest operating records in a network get the least attention, because nobody files a regional dispatch company or a claims administrator under "AI". The map below tests that against your own contacts.

| Term | Meaning on this page |
|---|---|
| Relationship holder | Whoever owns the relationship with the company: an engagement partner, a former colleague, an account manager |
| Path | The chain from you to the company's CEO, one name per link |
| Public clue | Something observable without asking the company, logged with source and date |
| Confirmed fact | Something the company told the relationship holder, dated |
| Eligible supplier | A company that passes the published fit facts and then SourceX's review; nobody on a fresh map is one yet |

## Why could your network's overlooked companies be part of AI's next growth wave?

Because AI attention and useful operating history sit in different places. Headlines follow model builders and AI startups; years of tickets, work orders, claims and project files sit inside ordinary businesses nobody has asked.

Epoch AI projects that, if current trends continue, the stock of public human-generated text could be fully used for training between 2026 and 2032, earlier if models are heavily overtrained. That projection concerns public text; it is no evidence that any company's private records are wanted, which SourceX judges only in review. It is a reason to look, never a reason a company qualifies.

### The overlooked-company field guide

Four pockets of a typical business network get skipped for predictable reasons. The fourth column is a hypothesis to raise in review, not a finding.

| Where it hides | Why it gets skipped | What its records might teach or test, if review agrees | The check that usually decides it |
|---|---|---|---|
| Long-standing clients, often owner-led | You see a client, not an owner of records | How staff scope, price and complete jobs, estimates beside actuals | Whether the files are the company's or its customers' |
| A former colleague's current employer | The contact is a peer, not a prospect | How cases move from intake to resolution in a ticketing, dispatch or claims system | Your former colleague is rarely the CEO and must ask |
| Vendors and suppliers you buy from | Money flows toward them, so nobody asks | How operations recover from order, quality and returns exceptions | Whether the account manager will take it to leadership |
| Companies that were acquired or wound down | They feel finished | Long, dated histories with known endings, if the archive survived | Who controls the records now; in a court-supervised wind-down the CEO alone may not be enough |

The [network opportunity finder](/tools/network-opportunity-finder) suggests sectors by role, and [referral income for former executives](/for/former-executive-referral-income) covers the alumni pocket from the other side.

## Which public clues suggest documented operating history, and what can they not prove?

Clues you can see without asking anyone for anything: how long the company has operated, which systems its job ads name, whether it publishes years of project history, and whether a dated migration, sale or closure is coming. None of them proves rights, record quality or demand.

| Public clue | What it may suggest | What it cannot tell you | Who can confirm it |
|---|---|---|---|
| A founding year on the company's own site | Years of records may exist | Whether old records survived system changes | The company, through the relationship holder |
| A job ad naming a ticketing, dispatch, claims or field-service platform | A named system holds the daily work | Who owns its data, or how far back it reaches | The system owner, through the holder |
| Case studies or project lists spanning a decade | Project histories were documented | Whether the files belong to the company or its clients | The CEO, once rights are raised |
| A dated public notice of a system migration, acquisition or closure | A real deadline for old records | Whether licensing suits the company's plans | The CEO |

Log each clue with its source and date. It becomes "confirmed" only when the company says so, and your reading of it sits in a separate "inference" column, so a guess never travels as a fact. Partners never ask for samples, exports or logins; the questions in the [metadata-only opportunity map](/guides/metadata-only-referral-opportunity-map) are the ceiling.

## Reachable owner or eligible supplier: what separates the two?

Reaching a company's owner tells you a conversation is possible. It does not tell you the company can license anything. Three checks separate the two, in this order.

1. **Permission.** When a colleague, client partner or former coworker holds the relationship, they decide whether to ask, and nobody goes around them. Then the CEO must know about the introduction and want it, because an approach the decision maker has not heard about is, in SourceX's guidance, usually not a fit.
2. **Ownership.** Does the company own the material, or can it license it without breaching client confidentiality, employee privacy or third-party contracts? And who says yes? That is the CEO, who authorizes a licensing discussion and signs on the company's behalf. A COO or system owner can champion the idea internally (weighed in [introducing through a sponsor or directly to the CEO](/compare/introduce-through-sponsor-or-directly-to-ceo)) but is not the authorizer. A parent and its subsidiaries count as one company for the reward cap.
3. **Mandate.** Yours, then SourceX's. Yours comes from engagement letters, employer policy and professional rules; SourceX tells licensed professionals to check their own rules on referral fees. For CPAs, the AICPA lists disclosing any commission or referral fees to the client among its Code's duties in financial planning work, and [referral fee rules by profession](/compare/referral-fee-rules-by-profession) gathers the others. SourceX's mandate is the program it is filling, so fit can change as programs change.

## How should an advisory firm build a referral opportunity map across its ecosystem?

Map relationship paths, not a contact list. An advisory firm's ecosystem has four lanes, and in each a different person's permission comes first. The owner-and-relationship opportunity map gives that person, the public clue, the confirmed facts and your inference separate columns.

- **Clients.** The engagement partner holds the path; the engagement letter and professional rules come before any reward.
- **Partners' own networks.** The partner whose contact it is decides; personal relationships are never pooled into a firm list.
- **Alumni.** Former staff hold the path to their own CEO and usually know which systems run the work.
- **Alliance firms.** Law firms, banks and software partners own their client relationships; ask the alliance contact, never the client. [Centers of influence](/guides/centers-of-influence) maps the professionals around an owner.

### The owner-and-relationship opportunity map (Illustrative)

Illustrative. One fictional advisory firm, generic companies, invented facts and dates, no outcome implied.

| # | Company | Lane and path to the CEO | Permission first from | Public clue (dated) | Confirmed by the company | Still inference | Next action |
|---|---|---|---|---|---|---|---|
| 1 | 160-person HVAC service contractor | Alumni: former analyst, now COO, then CEO | Former analyst | Job ad naming its field-service platform, 2026-10-02 | Work orders since 2009, own system (COO, 2026-10-14) | Whether service contracts limit reuse | COO asks the CEO |
| 2 | 240-person claims administrator | Client: engagement partner, then CEO | Engagement partner | Founded 2003, per its site | Nothing yet | Size, rights, records | Partner decides at the next review |
| 3 | 95-person regional distributor | Partner's network: tax partner, then founder-CEO | Tax partner (yes, 2026-10-08) | ERP migration announced for mid-2027 | 50+ full-time at peak; orders since 2011 (CEO, 2026-10-09) | Whether customer pricing is excluded | CEO wants a call: introduce |
| 4 | 300-person contact center outsourcer | Alliance: law firm partner, then CEO | Law firm partner | Case studies, 2015 to 2026 | Tickets sit in client systems (CEO, 2026-10-07) | None left | No fit: client-owned records |
| 5 | 400-person logistics company | No path into the firm's network | Nobody | Warehouse system replacement in trade press | Nothing | Everything | No contact: a clue is not a lead |

### A copy-ready row for your own map

| Column | What goes in it | Rule that keeps it honest |
|---|---|---|
| Company | A generic description | Client names never leave the firm |
| Lane and path to the CEO | One name per link | Say so if the chain stops below the CEO |
| Permission first from | The relationship holder | Their dated yes before anyone contacts the company |
| Public clue | One observable fact, source and date | No purchased lists or scraped contacts |
| Confirmed by the company | What was said, by whom, when | Metadata only: system, record type, years, rough count |
| Still inference | Your reading of the clue | Never upgraded without the company's word |
| Next action | Ask the holder, ask the CEO, revisit on a date, or drop | A dropped row records why |

**Method note.** Built from the published fit criteria (who qualifies, checked 2026-10-26) and the program terms (Version 2026-10-22.2), the map records permission and evidence, not demand, record quality or payout, which only SourceX's review assesses. No partner interviews or cohort results stand behind the lanes, and qualification rates for mapped companies are not yet published. The recommendation changes if your firm's rules bar referral rewards for clients (map only the other lanes) or if all three conversations below fail.

## Which three conversations come first, and what would challenge the thesis?

Choose three that each test a different assumption, so a week of conversations tells you whether your network holds overlooked companies at all.

| Conversation | Who you speak with | Assumption it tests | A result that would challenge it |
|---|---|---|---|
| 1. Strongest row (row 3) | The founder-CEO, via the tax partner | An owner with confirmed facts wants to explore licensing | The CEO declines once the process is explained |
| 2. Overlooked lane (row 1) | The former analyst, who asks his CEO | Alumni lanes hold records the firm never saw | The CEO will not engage, or the records are customers' |
| 3. Permission test (row 2) | The engagement partner, not the client | Holders in your firm will raise the topic at all | The partner will not raise it with any client |

If all three fail, the thesis is wrong for your network today: return to the 20-company screen above and revisit its "Not now" rows on a named trigger, such as a migration date or a headcount change. If only the first succeeds, your opportunity sits with owners you know well, not in the overlooked lanes.

### One relationship from clue to conditional outcome (Illustrative)

Follow map row 1 through each step.

1. **Discovery.** A partner sees the contractor's job ad naming its field-service platform and remembers that a former analyst now runs its operations. She holds only an inference: maybe years of work orders.
2. **Permission.** The former analyst agrees to raise it with the CEO and says plainly that he cannot say yes himself.
3. **Metadata.** Through the COO, the company confirms roughly 400,000 work orders since 2009 in a system it administers. Whether service contracts limit reuse goes to SourceX's rights review, not to the partner.
4. **Authorized introduction.** The CEO wants the conversation. The partner, who has signed the SourceX Channel Partner Agreement, sends her link and the company applies itself; credit goes to the first valid introduction within the 90-day window once it submits a verified application.
5. **Conditional outcome.** SourceX reviews the company against its current program. If it proceeds, the company signs one Master agreement, a Dataset Deal Confirmation fixes the dataset, net payout and terms, and SourceX pays the company. The reward is payable only after that supplier payment and the other conditions in the terms, and it falls due no later than 60 calendar days after whichever comes second: the dataset's completed handoff, or final written acceptance of the complete, validated dataset by the Acquisition Partner. Illustrative arithmetic: an Actual Supplier Payout of $240,000 means a $60,000 reward; from $400,000 upward the reward stays at $100,000.

**The no-fit case.** Row 4 had everything a contact list rewards: 300 staff, a decade of case studies and a law firm partner glad to ask. One answer from the CEO ended it: the tickets live in client systems under client contracts. A warm path does not cure ownership, and the row keeps the reason and date so nobody asks again next quarter.

> **Refer a company. Earn up to $100,000.** Default reward: 25% of Actual Supplier Payout, capped at $100,000 cumulatively per referred company. Eligibility and payment conditions apply, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

When a row on your map has the holder's yes and a CEO eager for the conversation, [sign up for a partner link](/register), hold the row up to [SourceX's qualification criteria](/who-qualifies) and go through [the referral terms](/terms); more partner playbooks are in [guides](/guides).

## Sources

- https://www.linkedin.com/help/linkedin/answer/a566336/export-connections-from-linkedin
- https://knowledge.hubspot.com/properties/hubspot-crm-default-company-properties
- https://refer.sourcex.si/terms
- https://epoch.ai/publications/will-we-run-out-of-data-limits-of-llm-scaling-based-on-human-generated-data
- https://refer.sourcex.si/
- https://refer.sourcex.si/who-qualifies
- https://www.aicpa-cima.com/resources/article/professional-responsibilities