The question matters because network programs are built around people: whom to brief, whom to invite to the dinner, whom to thank. If the roles that open conversations are not the ones a fund courts, the program spends its goodwill in the wrong place. The study is meant to show that before anyone redraws the program, and to stay silent until it can.
People make the asks; roles are what the study records
A chain might run from a fund partner to a limited partner, from the LP to the owner of a company the LP backed, and end with that owner agreeing to a call. The study records the chain as three role codes and two links, never the names. Roles protect the people involved, and they travel: another fund has different people in the same positions, so a role map can be compared and checked.
Every chain ends, if it ends well, at the company's CEO. Under the program the CEO authorizes a licensing discussion, so an inside executive who hears the idea is a link in the chain, not its end. For a company that has been wound down or is in a formal process, the CEO may not be enough; the end node is whoever controls the estate, and those chains are reported apart.
The role codebook
Twelve role codes cover the positions a partner's ask is likely to pass through. Two coders apply them independently to a sample of links and report how often they agree.
| Code | Role | Where it sits relative to the company |
|---|---|---|
| R1 | Fund partner or platform lead | Investor side, usually where chains start |
| R2 | Portfolio founder-CEO | Peer reached through the fund |
| R3 | Board director | Governance seat at the company |
| R4 | LP or family office principal | Owner or investor network |
| R5 | Outside lawyer | Adviser to the company |
| R6 | Outside accountant or fractional CFO | Adviser to the company |
| R7 | Banker or broker | Transaction adviser |
| R8 | Peer CEO at another company | Peer outside the fund's portfolio |
| R9 | Executive inside the company, such as the CFO or COO | Inside, holds records, does not decide |
| R10 | Customer or supplier contact | Commercial counterparty |
| R11 | Association or community leader | Hosts a network the company belongs to |
| R12 | Consultant or implementation partner | Serves the company under an engagement |
Two rival predictions from network research
Network research offers a reason to doubt the obvious answer. Rajkumar and colleagues, writing in Science in 2022, report randomized experiments on LinkedIn's People You May Know feature that varied how many weak ties more than 20 million people had over five years; in their results, weaker ties increased job transmission only up to a point, after which the gains fell off (paraphrased from the published summary). Job moves are not licensing conversations, so the finding is a reason to test, not a forecast.
| Prediction | What it expects | What would support it |
|---|---|---|
| H1, closeness | Strong ties unlock most: portfolio founders, board directors and LP principals in frequent contact with the CEO | Unlock ratios rise steadily with contact frequency and years known |
| H2, moderate distance | Peer CEOs, association leaders and other looser ties unlock at least as often as the closest ties | Unlock ratios peak in the middle tie-strength band |
Frame, consent and the fields logged for every link
The entry window is one calendar quarter. Participating partners log every ask they start, including the ones that went nowhere, and the cutoff for outcomes falls 90 days after the window closes. Intermediaries are told their role code will be logged without their name, and onward links are recorded either by the intermediary or by what they tell the partner, marked as which. Companies appear as a size band and a sector. No records, documents or system details are requested at any point.
The fields for each link:
- Chain and link number, so a chain can be rebuilt in order.
- From-role and to-role, using the codebook, with CEO or estate controller as the possible end.
- Date of the ask.
- Tie strength from the asker's side: contact weekly, monthly, or yearly or less; known for under one year, one to five years, or more than five.
- Link outcome: passed on, declined to pass on, or no answer.
- Chain outcome at cutoff: CEO agreed to a conversation, CEO declined, dead end, or still open.
The role-to-conversation network analysis
Five measures come out of the pooled chains, and each is reported with its count. A role with fewer than ten chains is pooled into "other" rather than ranked on a handful of cases. The bridging measure borrows a standard definition: NetworkX describes the betweenness centrality of a node as "the sum of the fraction of all-pairs shortest paths that pass through v". On a role graph it shows which positions sit between the fund and the CEOs it cannot reach directly. It says nothing about the value of those positions.
| Measure | Definition | What it answers |
|---|---|---|
| Unlock ratio by role | Chains through a role that end with the CEO agreeing, divided by all chains through that role; open chains give a lower and an upper bound | Which positions sit on successful chains |
| Dead-end share by role | Asks a role received that went no further, divided by all asks it received | Where chains die |
| Hops to the CEO | Links in each chain that reached the CEO, grouped by the first role asked | Whether long chains are worth starting |
| Bridging score | Betweenness centrality on the pooled role graph | Which positions connect otherwise separate parts of the network |
| Unlock ratio by tie strength | The first measure split by contact frequency and years known | Which of H1 and H2 the data favor |
When chains from different partners reach the same company, the study counts that convergence and nothing more. Which introduction, if any, is credited is SourceX's decision from dated records in its duplicate review.
Three chains as they would be logged (Illustrative)
Illustrative. Generic companies, invented facts, no outcomes implied.
| Chain | Links | Tie strength at the last link | Outcome at cutoff | How it is coded |
|---|---|---|---|---|
| A | R1 to R4 to the CEO of a 180-person distributor | Monthly contact, known over five years | CEO agreed | Two hops; counts toward R1 and R4 unlock ratios |
| B | R12 to R9, the CFO of a 350-person logistics firm | Weekly contact, one to five years | CFO never raised it; no answer | Dead end at R9 |
| C | R1 to R2 to R8 to the CEO of a 400-person staffing firm | Yearly contact, under one year | CEO declined | Three hops; reached but declined, in R8's denominator |
Patterns that would not mean what they seem
Partners choose whom to ask, so the study is observational. If LP principals show a high unlock ratio, it may be because partners ask LPs only about companies the LP owns, not because the role itself opens doors. Two comparisons narrow that gap: roles compared within the same partner, and within the same tie-strength band. Neither turns the study into an experiment, and the report says so beside every comparison.
A high bridging score marks a position that connects parts of the network. It does not show that the person in that position should be paid, thanked or asked more often. A chain that dies at an inside executive is also not a judgment of that executive: under the program, only the CEO can open the discussion.
Results that would redraw the program
- If unlock ratios overlap across roles once tie strength is held equal, then relationships matter and titles do not, and a program should map ties, not job titles.
- If dead ends cluster at inside executives, asking for the CEO first is supported, as the program's authorization rule already suggests.
- If H2 holds, peer-CEO and association routes deserve more of a platform team's time than the fund's closest relationships.
- If chains of three or more links almost never reach the CEO, long chains are not worth starting.
How the map will be published
Nothing is drawn until the cutoff has passed. The published report will carry the chain flow with counts at every stage, the role graph with node size set by chains and line width by asks, the five measures with their bases, counterexamples that cut against the main pattern, and the coder agreement rate. No partner, intermediary or company is identifiable, and no output is broken down by partner.
Method note. The codebook, fields and measures rest on SourceX's published authorization rule and fit criteria (the terms and the who-qualifies page, checked 2026-10-28), the tie-strength question raised by the Science study and NetworkX's definition of betweenness. The study measures which roles carry asks and where chains end. It does not measure SourceX's review decisions, demand for any company's records, licenses or rewards.
Two neighbouring studies answer different questions. The sponsor-reach study asks whether one introducer's path reached the CEO, by path type; the introducer-chain ledger records what each person in one chain knows and does. This study pools chains across a network. For acting on relationships today, the five-ring map sorts companies by who owns each CEO relationship, and SourceX's own standard for citing results is in the partner results research guide.
Risks, limits and what the study cannot see
- Partial chains. A partner sees their own link and what they are told about the next; hops nobody reports are invisible.
- Recall. Tie strength is the asker's own estimate.
- Who volunteers. Partners who log every ask are likely the more careful ones.
- Small numbers. One quarter may leave most roles pooled into "other". The report says so rather than guessing.
- Moving terms. SourceX versions its program terms; a change during the window is dated and the cohort split at it.
- Distance from outcomes. A CEO who agrees to talk has not applied, been accepted or licensed anything.
Before any chain starts
The company at the end of a chain should plausibly fit. Its full-time headcount should have reached 50 at its peak, leaving contractors out; it should hold an operating history recorded in its own systems and be free to license that history; and its CEO should already know about the idea. 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 70% of the data is in English. SourceX credits one partner per company, from dated records, once the company submits a verified application inside the 90-day window, and a click on a link decides nothing. Whoever submits signs the Channel Partner Agreement first.
Next step
Refer a company. No cap on rewards. Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.
When a chain in your own network ends at a CEO who wants the conversation, register as a partner, set that company beside who qualifies and go through the program terms ahead of any submission; other research designs are collected in the guides hub.
The study maps relationships for referral work; whether SourceX itself is a sound investment is a separate question, and materials for that are available on request through the SourceX investor overview.
This article is for information only and is not an offer to sell or a solicitation of an offer to buy securities. Any offering by SourceX is made only to verified accredited investors through its offering documents.