# Referral cohort report across criteria versions | SourceX

Canonical: https://refer.sourcex.si/guides/a-cohort-report-for-referrals-made-under-different-criteria
Last updated: 2026-11-05

Never pool introductions made under different rules into one rate. Tag every introduction with the criteria in force on the day it was made: your fund's own referral mandate, the program-terms version and, where SourceX tells you, the program the company was reviewed against. Count unique companies per cohort, report every stage including cases still open, and compare cohorts only at the same age since introduction. With cohorts this small, publish counts with denominators, such as 3 of 9, never bare percentages.

A pooled rate answers no question anyone at the fund is asking. The question is usually narrower: did tightening our screen in the second quarter mean fewer wasted introductions? That needs two cohorts, one rule set each.

## Three layers of criteria that change under a platform team

A fund's referral record is shaped by three sets of rules, and each changes on its own schedule.

| Layer | Who changes it | Where its version lives | What a change can move in your report |
|---|---|---|---|
| The fund's referral mandate | Your platform team | Your own dated mandate memo | Which companies get introduced at all |
| The program terms | SourceX | The version number and change log at the foot of the terms | Credit, duplicates, reward calculation and payment timing |
| The program a company is reviewed against | SourceX, per data program | Only what SourceX shares with you | Whether a company fits now or later |

**Your mandate.** An Illustrative example: mandate v1 introduced any portfolio company whose founder asked; mandate v2 required 50+ full-time employees at peak, at least five years of the company's own records and the CEO's recorded consent before submission. Date the memo the day it takes effect.

**The terms.** The current terms are Version 2026-10-22.2, since replaced by version 2026-11-04, and their change log records five recent versions: from 2026-10-12 the reward calculation of the greater of $50,000 plus 5% and 15% of Actual Supplier Payout applies to all referrals, including earlier ones; from 2026-10-19 partners sign the Channel Partner Agreement; from 2026-10-21 section 3 says what counts as an existing relationship, which now includes any earlier lead, account or email correspondence with SourceX; from 2026-10-22 section 5 states the payment deadline and conditions; and from 2026-10-22.2 section 4 defines Actual Supplier Payout. Only the first entry says it reaches back to earlier referrals. For the rest, record the version in force on the introduction date and let SourceX's review apply it; the terms also say signed agreements are never silently replaced.

**The program.** SourceX judges each company against the criteria of a specific data program, not one universal minimum, so a company declined under one program could still suit another one later. A decline is therefore dated, not permanent, and a cohort report should say so.

## Stages, populations and one cutoff date

Define each stage before you count anything, and use the records SourceX actually shares with you. Partners follow their referrals through generic stages or reference codes, so a report that needs Acquisition Partner names or prices cannot be built and should not be attempted.

| Term | Definition used in this report | Record that proves it |
|---|---|---|
| Entry event | A written introduction made with the CEO's recorded consent | The introduction record: company, domain, partner, opportunity, method and timestamps |
| Population | Unique companies, one row each, by first introduction date | Your deduplicated log |
| Entry window | The introduction dates a cohort covers | Your mandate memo dates |
| Credited | The company submitted a verified application within the 90-day attribution window | SourceX status update |
| Attribution outcome | Duplicate, existing relationship or no application in time; not a judgment on the company | SourceX status update |
| Outcome | Declined with a reason, withdrew, or reached a reward status | SourceX status update; reward statuses Estimated, Eligible, Approved, Processing, Paid |
| Age | Days since the company's introduction | Your log |
| Cutoff | One observation date applied to every cohort | Stated at the top of the report |

## The criteria-version cohort comparison (Illustrative)

Illustrative and fictional: invented counts for an imaginary fund, built to show the format. They are not SourceX data and not any fund's results.

| Line | Cohort A, mandate v1 | Cohort B, mandate v2 |
|---|---|---|
| Entry window | First 90 days of participation | The following 90 days |
| Terms version | Recorded per introduction | Recorded per introduction |
| Log entries | 14 | 9 |
| Unique companies introduced | 11 | 8 |
| Screened out by the mandate before any introduction | Not recorded; v1 had no screen | 6 |
| Attribution outcomes | 1 | 1 |
| No verified application within 90 days | 1 | 0 |
| Credited companies | 9 | 7 |
| Credited companies that had reached day 60 by the cutoff | 9 | 5 |
| Too young to compare at day 60 | 0 | 2 |
| Declined at eligibility by day 60 | 3 of 9 | 1 of 5 |
| Declined on rights or data fit by day 60 | 1 of 9 | 0 of 5 |
| Withdrew by day 60 | 1 of 9 | 0 of 5 |
| Still open at day 60 | 4 of 9 | 4 of 5 |
| Rewards paid at the cutoff | None | None |

Every column reconciles: 11 companies less 1 attribution outcome and 1 missing application leaves 9 credited in cohort A, and 3, 1, 1 and 4 add back to 9. In cohort B, 8 less 1 leaves 7 credited, of which only 5 are old enough to compare.

**What the table supports.** Under mandate v2, 1 of 5 credited companies failed eligibility by day 60, against 3 of 9 under v1. That is consistent with a stricter screen and far too small to call a difference. Cohort B also removed 6 companies before introduction that v1 would have sent; those six are part of what the mandate did, and a fair reading tracks them. Read the cohorts again at day 120, once B reaches it.

**What it does not support.** Any statement about reward value, any rate with a percent sign, or any claim that v2 "converts better". Nothing has resolved to a payment in either cohort.

**Method note.** The comparison uses three records: the fund's dated introduction log, SourceX's status updates and the terms version in force on each introduction date, all read at one cutoff. Age is measured per company. It does not measure demand for any company's records, payout or reward amounts, and it cannot see which data program a company was assessed against unless SourceX shared it. The step-by-step coding of decline reasons is in the [rejection reasons study](/guides/data-referral-rejection-reasons-study), and the choice of headline denominator in the [conversion benchmark design](/guides/data-referral-conversion-benchmark).

## Reconcile unique companies and open cases before comparing

1. **First introduction decides the cohort.** A company introduced under v1 and again under v2 stays in cohort A. The second entry is a duplicate row in your log.
2. **Attribution outcomes are not declines.** A company that already had an application, relationship or earlier valid introduction with SourceX is reviewed as a duplicate; report it on its own line. The page on [a company already known to SourceX](/questions/what-happens-if-the-company-was-already-known-to-sourcex) explains why.
3. **Group companies once.** A parent and its subsidiaries share one reward cap. Whether a sister company's earlier contact, or another partner's introduction of a sibling, affects credit is decided by SourceX in duplicate review; your report records the outcome and does not predict it.
4. **Open cases stay in the denominator.** At every cutoff, each company is declined, withdrew, attribution outcome, reward status reached, or open.
5. **Roll forward.** Each report states how many open cases moved since the last one and where they went, so nobody can quietly drop a stale row.
6. **Re-entry is a note, not a new success.** A company declined under one program and reviewed later under another keeps its original row with a dated note.

## Report the mandate lesson, not a projected payout

A cohort report exists to improve the next mandate. Five lines carry that lesson without forecasting anything:

- screen-outs before introduction, with the fund's own reasons;
- eligibility declines per credited company at equal age;
- median days from introduction to verified application, once more than half of a cohort has applied;
- attribution outcomes, separately;
- open cases, with their ages.

Leave out reward totals built from Estimated statuses; the terms call such figures projections, not promises. Do not annualize a quarter's paid rewards. The visibility problem is not unique to funds: in PartnerStack's 2026 survey of 100 senior leaders at B2B SaaS companies with $50M+ revenue, 14% named lack of visibility into partner impact as the biggest blocker to partner-driven revenue, behind team alignment at 37%. A report that states its own definitions removes one source of that blindness.

## Where a criteria-version comparison misleads

- **Small numbers.** With single-digit cohorts, one company moves a comparison by double digits in percentage terms. Counts and denominators only.
- **Confounding.** A new mandate often arrives with a new platform lead, a different slice of the portfolio or partners who have simply learned the fit tests. The comparison cannot separate those.
- **Moving program demand.** SourceX checks relevance against its current data needs, and requirements vary by program, so a decline rate can shift without any change on the fund's side.
- **Hidden review criteria.** The fund sees generic stages, not the program each company was assessed against.
- **Retroactive terms.** A change the terms apply to earlier referrals, such as the 2026-10-12 reward calculation, can alter an older cohort's figures after the fact; note it in the report.
- **Censoring.** Younger cohorts have more open cases by construction. Compare at equal age or not at all.

## Which companies may enter a cohort

A cohort should only contain companies that could plausibly fit. In practice, each one needs a peak headcount of 50+ full-time employees, not counting contractors; an operating history it can document; the right to license its records; and a CEO who is aware of the conversation and authorizes the approach. 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. Funding status never establishes eligibility.

Platform teams gather only metadata: record types, rough counts, years covered, systems and their owners. They never collect files, samples or credentials. What stays unknown until SourceX reviews the company: demand for its records, contract limits on reuse, and what any license would cover. Whoever submits must first sign the SourceX Channel Partner Agreement; a partner who signed it back when it was titled the Channel Partner Non-Circumvention and Non-Disclosure Agreement is still held to that version as signed. Platform leads setting up a portfolio-wide program can start from the [venture platform team overview](/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.

Before your next mandate takes effect, date it, [register for a partner link](/register), screen candidates against [who qualifies](/who-qualifies) and note the [program terms](/terms) version in force, so the next cohort can be compared honestly with this one.

A fund evaluating SourceX itself as a company is making a different decision from running a referral cohort; investor materials are available on request through the [SourceX 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://refer.sourcex.si/who-qualifies
- https://refer.sourcex.si/rewards
- https://sourcex.si/documents/partner/SourceX_Channel_Partner_Agreement_Sample.pdf
- https://partnerstack.com/resources/research-lab/the-state-of-partnerships-in-gtm-2026
- https://sourcex.si/investors
