# Repeat referral partner behavior, measured | SourceX

Canonical: https://refer.sourcex.si/guides/how-to-measure-repeated-introductions-from-the-same-network
Last updated: 2026-11-05

Count returning people, not returning companies. A network repeats when the same introducer, the person who owns a CEO relationship, brings a second new company that reaches a verified application in a later quarter. Re-sending a company SourceX already knows, a renewal by a company introduced earlier and an introduction of a sister entity are separate events with separate rules. Group introducers by the quarter of their first verified application, follow each group quarter by quarter, and print the denominator, the open cases and the quarters not yet observed in every row.

A fund COO usually asks this after a good first quarter: was it a burst from three enthusiastic people, or the start of a habit? Answering it needs a report that double counting cannot flatter and that does not punish recent cohorts for having had less time.

## Four look-alikes and the one return that counts

Five things can happen the second time a name from your network appears in the referral records. Only the first one measures repeat introducer behavior.

| Second appearance | Example | What the program says | Counted as a return? |
|---|---|---|---|
| Same introducer, new company, verified application | A portfolio founder who introduced a distributor in Q1 introduces the CEO of a 300-person insurer in Q3 | The first valid introduction of a new company is credited when that company submits a verified application | Yes |
| Same company, introduced again | A colleague re-sends the distributor | A company with an existing application, relationship or earlier valid introduction counts as a duplicate, so no later partner receives credit | No; it stays one company row |
| Same company, second license or renewal | The distributor signs a second Confirmation | Repeat licenses and renewals agreed during the agreement's term or protection period stay attributed; a further qualifying transaction first agreed within 12 months of the first Closing pays 15% of its Actual Supplier Payout | No; it belongs in the reward ledger |
| Sister entity of an earlier company | Another subsidiary in the distributor's group | Whether a sister entity is a new referred company, with its own once-per-company USD 50,000 floor, and how the group's earlier contact affects credit is decided by SourceX in duplicate review | Pending until review decides |
| Same company after a lapsed window | No application arrived within 90 days, and the introducer tries again months later | Whether a later attempt can be credited depends on what SourceX's records already hold, settled in duplicate review | Pending until review decides |

The second row matters more than it looks. Section 3 of the terms treats any earlier lead, account or email correspondence between SourceX and a company as an existing relationship, however it began. A network that keeps surfacing the same well-known companies will log plenty of activity and almost no returns.

## Who is in a cohort, which event counts and when the clock stops

Write five definitions at the top of the report and keep them fixed for the year.

- **Unit: the introducer.** The person who owns the CEO relationship and made the ask. SourceX credits one partner account per company, and in a fund the submitting account is not always the person who knew the CEO. Record both; measure the relationship owner.
- **Cohort: the quarter of a person's first verified application.** Not the quarter they registered or sent a first email. Attribution locks when the application is submitted, so it is the earliest date SourceX's records fix.
- **Return event: another new company reaching a verified application** in a later quarter, introduced by the same person. Two companies in one quarter are one returning introducer and two companies.
- **Open case:** a later introduction still inside its 90-day window, or a company in duplicate review. An introducer whose only later company is open is "unresolved", never "not returned".
- **Observation cutoff:** the date statuses were read. A cell that falls after the cutoff is "not yet observable", never zero.

One structural fact distorts long-run retention if you ignore it. Partners sign the SourceX Channel Partner Agreement before submitting introductions, and it runs for one year and renews only with fresh acceptance and signatures. An introducer whose agreement has run out cannot submit until it is renewed, so a quiet fifth quarter can be an enrollment gap rather than a cooled relationship. Give "agreement lapsed" its own line.

Every return still has to clear the published fit facts on its own: 50+ full-time employees at peak with no contractors counted, a documented operating history, rights to license, and location. 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. The CEO authorizes a licensing discussion and signs for the company, and introducers describe records at metadata level without handling them.

## The partner cohort retention report

The report is a triangle. Each row is a cohort, each column a quarter after that cohort's first verified application, and each cell the number of the cohort's introducers who brought another new company to a verified application in that quarter, over the cohort size. Two closing columns summarize returns and open cases at the cutoff.

Illustrative: an invented fund network of 17 introducers across the five quarters of one program year, statuses read at the end of quarter 5.

| Cohort | Introducers | Quarter +1 | Quarter +2 | Quarter +3 | Quarter +4 | Returned by cutoff | Unresolved at cutoff |
|---|---|---|---|---|---|---|---|
| Q1 | 4 | 1 of 4 | 1 of 4 | 0 of 4 | 1 of 4 | 2 of 4 | 1 |
| Q2 | 3 | 0 of 3 | 1 of 3 | 1 of 3 | Not yet observable | 1 of 3 | 0 |
| Q3 | 5 | 1 of 5 | 0 of 5 | Not yet observable | Not yet observable | 1 of 5 | 2 |
| Q4 | 2 | 0 of 2 | Not yet observable | Not yet observable | Not yet observable | 0 of 2 | 1 |
| Q5 | 3 | Not yet observable | Not yet observable | Not yet observable | Not yet observable | Not yet observable | 0 |

How to read it:

- **The headline is 4 of 14, not 4 of 17.** Only the 14 introducers in cohorts Q1 to Q4 have had a full quarter in which to return. The Q5 row is right-censored: NIST's engineering statistics handbook uses that term for data cut off at a fixed time, where event times beyond the cutoff are unobserved. Unobserved is not the same as absent. Counting those three in the denominator would understate retention; deleting their row would hide that they exist.
- **Unresolved gets its own column.** Four introducers have a later company in its 90-day window or in duplicate review. If all four confirm, the figure becomes 8 of 14; if none does, it stays at 4 of 14. Print the range, not a midpoint.
- **Read the diagonals as well as the rows.** Quarter 5 sits in the Q1 row's fourth column, the Q2 row's third, the Q3 row's second and the Q4 row's first. Something that changed program-wide in that quarter, such as a SourceX program closing, shows up down that diagonal.
- **Check concentration.** The six return events came from four people, and two of those people produced four of the six. A network that leans on one or two prolific introducers is fragile, however tidy its triangle looks.

At this size, keep cells as counts. "1 of 5" tells the reader how much weight a cell can bear; "20%" hides it.

**Methods note.** The design applies the attribution and duplicate rules in the program terms (version 2026-10-22.2, since replaced by version 2026-11-04), the term and repeat-transaction rules in the Channel Partner Agreement sample and NIST's description of right-censored data, all checked 2026-10-26. It measures whether introducers come back with new companies. It does not measure demand for anyone's records, acceptance or payouts, and it contains no SourceX figures; SourceX has published none at network level.

## Tie every return to one company row

A retention triangle is only as good as the company list underneath it. Before the report goes anywhere, reconcile introductions to companies and companies to people.

Illustrative, the same network, quarters 1 to 5:

| Line | Count | Where it goes |
|---|---|---|
| Introductions recorded in writing | 31 | Starting total |
| New companies at verified application, first per introducer | 17 | Forms the cohorts |
| New companies at verified application, from a returning introducer | 6 | Return events in the triangle |
| Open: inside 90 days or in duplicate review, one of them a sister entity | 4 | Unresolved column |
| Company already known to SourceX | 2 | Excluded; reported as an outcome |
| Same company introduced by a second colleague | 1 | Counted once, under the first introducer |
| Window closed with no application | 1 | Excluded; reported as lapsed |

The six outcome lines add up to 31, so no introduction has gone missing. Then reconcile the people: 17 introducers form the cohorts, and of the 14 with a full quarter of follow-up, 4 returned, 4 are unresolved and 6 have not returned. One of those six let the agreement run out in its twelfth month and did not renew, which belongs on the "agreement lapsed" line, not in a story about the network cooling.

A Q1 company also renewed its license in the period. It is real business and appears in neither table, because it is a further transaction with the same company, paying 15% of its Actual Supplier Payout when first agreed within 12 months of the first Closing; it belongs to that company's reward ledger.

Partners follow their own referrals through generic stages or reference codes, which is all this reconciliation needs. Keep the reference code beside each company row so a later status change can be matched without names. The [repeat partner worksheet](/resources/a-repeat-partner-activity-and-outcome-worksheet) gives a per-company log to feed it, and [tracking by network segment](/guides/how-to-track-accepted-introductions-by-professional-network-segment) adds the part of the network each introducer comes from.

## A returning introducer is a habit, not a revenue line

Retention measures behavior. A fund should resist turning it into money.

- **No reward run rate.** A company's reward accrues only as SourceX pays that company, and before supplier payment any reward shown is an estimate. Four returning introducers do not imply four rewards a year.
- **A return can be worth nothing.** A second company can be declined, can turn out to be known to SourceX already, or can fit a program that is not open this quarter. SourceX applies each program's own criteria, so a return that misses one program may still be a fit for another later on.
- **A second license is one company's business, not network growth.** If an earlier company licenses again, a further qualifying transaction first agreed within 12 months of its first Closing pays 15% of its Actual Supplier Payout, and the USD 50,000 floor does not apply a second time. The page on [repeat business and the Repeat Window](/questions/referral-fee-on-repeat-business) sets out the per-company rules.

What the report can support is a decision about effort: which introducers to brief again, which parts of the network keep producing companies that pass the fit facts, and whether the habit is spreading or resting on two people. Whether rising retention is also turning into better-fitting companies cohort after cohort is a separate test, set out in [six cohort signals of a compounding network](/guides/how-to-recognize-a-partner-network-that-is-starting-to-compound). The [illustrative repeat partner case](/resources/repeat-referral-partner) shows the consent-first routine that a healthy return rate should rest on.

## Where a retention triangle misleads

- **Tiny cohorts.** With two to five introducers per row, one person moves a cell by 20 to 50 points. Read several quarters together and keep counts.
- **Forgotten censoring.** Recent cohorts always look worse because they have had less time. Never compare a Q4 row's "returned by cutoff" with a Q1 row's without saying how many quarters each has had.
- **Review lag.** Open cases can stay unresolved for weeks. Re-issue the report after duplicate review decisions, with the new cutoff date printed.
- **Volume pressure.** Celebrating returns can nudge introducers toward companies whose CEO has not agreed to a conversation. An introduction made without the decision maker's knowledge is usually not a fit, and it spends the relationship.
- **Corporate groups.** Sister companies are where repeats are most often miscounted. The USD 50,000 floor applies once per referred company, so whether a sister company is a new company or further business of one already introduced, and whether the group's earlier contact changes its credit, is for SourceX to decide in duplicate review; keep those rows pending until it does. [Multiple subsidiaries in attribution](/guides/how-to-handle-multiple-subsidiaries-in-referral-attribution) covers the group cases.
- **Program changes.** Retention can fall because SourceX's programs changed, not because the network tired. Note any known program change beside the diagonal it touches.

The fund-side overview of the program is on the [venture platform page](/for/venture-platform-teams).

## 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 the triangle shows which introducers come back with companies that pass the fit facts, make sure each of them has [registered](/register) and signed, check their next company against [who qualifies](/who-qualifies), and keep the [program terms](/terms) on duplicates beside the report.

Counting how often your own network returns is referral work; judging SourceX as a business is a different question, and SourceX shares its investor materials on request at the [investor overview](https://sourcex.si/investors).

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.

## Sources

- https://refer.sourcex.si/terms
- https://sourcex.si/documents/partner/SourceX_Channel_Partner_Agreement_Sample.pdf
- https://refer.sourcex.si/who-qualifies
- https://www.itl.nist.gov/div898/handbook/apr/section1/apr131.htm
- https://sourcex.si/investors
