A referral log is a list of names someone typed. A pipeline report is a claim about companies. The gap between the two is where inflation lives, and in a fund it usually comes from good intentions: two colleagues log the same founder's company, a platform analyst counts a parent and its subsidiaries separately, a dashboard estimate is summed as "potential". (For the other meaning of pipeline, the route records take from a company's systems to a buyer, see what a data pipeline is.)
Seven ways a referral pipeline inflates
Each trap below is common, each makes the channel look busier than it is, and each has a rule that removes it.
| Trap | How it inflates the report | Rule that removes it |
|---|---|---|
| Entry counting | The same company appears twice because two people logged it, or once under an old name | One row per company, keyed on legal name and primary domain |
| Group splitting | A parent and its subsidiaries count as three wins | One reporting unit per group; entity count in a note |
| Known-to-SourceX companies | A company with an earlier lead, account or email with SourceX is reported as pipeline | Its own row, "not credited", never in the active count |
| Stale introductions | An introduction with no verified application inside 90 days stays "active" | Mark it lapsed on the day the window closes |
| Estimates as value | Dashboard estimates, or names multiplied by the reward floor, are summed into a "potential" figure | Counts only until a reward is paid |
| Cumulative as periodic | A running total since launch is presented as this quarter's activity | Cohorts by introduction date, measured at a stated cutoff |
| Small-number ratios | "Two of three applied" becomes 67%, then a run rate | Counts with denominators; no percentages under 20 cases, no annualizing |
Two of these rules come straight from the program terms. Each company yields credit for one partner only, the partner behind the first valid introduction of a company new to SourceX, and that credit depends on the company submitting a verified application inside the 90-day attribution window. And section 3 classes a company as a duplicate when SourceX already has any earlier lead, account or email correspondence with it, and a later partner gets no credit for a duplicate. A report that ignores either rule is counting companies the program will never credit.
Stages, populations and the cutoff date
Fix three things before counting anything.
The population is every unique company introduced during the cohort window, here the third quarter, by anyone reporting to the fund. Logged names that never became an introduction are counted on the way in, then removed.
The cutoff is the date the report reads statuses, here October 15. Every company sits at the furthest stage it reached by that date, and later events wait for the next report.
The stages, with the evidence a fund can actually hold:
| Stage | Definition | Evidence |
|---|---|---|
| Logged | A name entered in the fund's tracker | Your tracker |
| Introduced | A written introduction sent with the CEO's agreement, by someone who has signed the Channel Partner Agreement | The dated introduction record |
| Credited | A verified application within 90 days, and the first valid introduction | SourceX status, by reference code |
| In review | SourceX is assessing the company | SourceX's generic stage |
| Declined | Review ended without a deal | SourceX's recorded outcome |
| Signed | The company says it signed its Master agreement and a Confirmation | The company's word |
| Supplier paid | SourceX has paid the company | Reward status moves past Estimated |
| Reward paid | The reward reached the partner | Status Paid, and your bank record |
The partner agreement helps here. It requires every introduction record to identify the company, its domain, the partner, the opportunity, the introduction method and timestamps, which is exactly the key a deduplication needs. It also limits what you will see: generic stages and reference codes, not the Acquisition Partner's name, price or acceptance evidence, and never another partner's records. Where a stage is not visible to you, the cell says "not known", which is a fact, not a guess. Where a partner can see the current stage covers the dashboard side.
The deduplicated pipeline waterfall (Illustrative)
Illustrative. One fund's third-quarter log, read on October 15. Generic companies, invented counts, no outcomes implied for any real company.
| Step | Count | Removed | Rule applied | Evidence |
|---|---|---|---|---|
| Entries logged in the quarter | 31 | Starting count | Raw tracker rows | Tracker export, dated |
| Less duplicate entries | 25 | 6 | Same legal name or domain, logged twice | Name and domain match |
| Less entities folded into a group | 22 | 3 | Parent and subsidiaries become one reporting unit | The CEO's description of the structure |
| Less names never introduced | 17 | 5 | No written introduction with the CEO's agreement | Absence of an introduction record |
| Introduced companies | 17 | 14 in all | The report's denominator | Introduction records |
Then every one of the 17 lands in exactly one outcome row:
| Outcome on October 15 | Companies | What resolves it |
|---|---|---|
| Not credited: earlier relationship with SourceX | 2 | Resolved |
| Lapsed: no verified application within 90 days | 1 | Resolved |
| Unresolved: window still open, no application yet | 4 | The application, or the window closing on its dated day |
| Unresolved: in duplicate review | 1 | SourceX's decision |
| Credited | 9 | Moves on through review |
| Total | 17 | Must equal introduced companies |
And the 9 credited companies, by furthest stage: 6 in review, 1 declined, 2 signed by the company's account, 0 supplier paid, 0 rewards paid.
Set that beside the version that tends to circulate: "31 opportunities, 9 deals in progress, 2 closed, $1.55 million potential." Every word of it can be defended by a spreadsheet, and almost none of it describes companies. The 31 includes duplicates and names nobody introduced; "closed" means two companies say they signed, which earns nothing on its own; and $1.55 million is 31 names multiplied by the USD 50,000 floor on a first qualifying transaction, a number this pipeline cannot support, because the floor is paid only once a company's first qualifying transaction is paid, and most rows will pay nothing. The deduplicated report says 17 companies introduced, 9 credited, 5 unresolved and $0 paid, and each figure survives a question.
Every company in one row, every open case rolled forward
Three checks keep the report honest from one quarter to the next.
- Each company appears in exactly one outcome row, and the rows sum to the denominator. If they do not, a duplicate slipped through or a company fell out.
- Every unresolved row names its resolver. A window that is still open carries the date it closes; a duplicate review carries "SourceX". Nothing stays unresolved without a reason.
- Unresolved cases roll forward, never drop out. The next report opens with them.
| Third-quarter cohort | October 15 | January 15 |
|---|---|---|
| Introduced companies | 17 | 17 |
| Credited | 9 | 11 |
| Not credited | 2 | 3 |
| Lapsed | 1 | 3 |
| Unresolved | 5 | 0 |
| Of credited: in review, declined, signed, supplier paid | 6, 1, 2, 0 | 4, 2, 3, 2 |
| Rewards paid | 0 | 0 |
Between the two dates, two open windows produced verified applications and two closed without one, and the duplicate review went against the fund. The denominator never moved. That is the property that makes cohorts comparable: the third quarter is always 17 companies, however its outcomes develop. Join your rows to SourceX's statuses by reference code rather than by company name, which avoids the errors that come with renamed or re-registered companies; when a company registers twice and preventing duplicate claims cover the upstream fixes.
A one-page quarterly report, with income left at what was paid
A quarterly referral report fits on one page:
- Counts with denominators. "17 companies introduced; 9 of 17 credited; 5 of 17 unresolved: 4 windows closing by December 29 at the latest, 1 in SourceX's duplicate review."
- What moved since the last report, by cohort, never pooled across quarters to make a ratio look sturdier.
- Why companies fell out, in your own words where SourceX shares a reason; the rejection reasons study offers a taxonomy.
- What the fund will change. In the Illustrative quarter, six of the fourteen removals were duplicate entries and two companies were already known to SourceX, so the fix is upstream: one shared tracker and a check before anyone writes to a CEO.
- Income, as paid rewards only. Estimates do not appear, and nothing is multiplied by four.
Attribution method is a reason to be strict. PartnerStack's State of Partnerships in GTM 2026 research found that 42% of surveyed B2B SaaS companies use multi-touch attribution for partner revenue, 31% first-touch, 19% last-touch and 8% other or no tracking. A fund that borrows its CRM's touch model will count differently from the program. For SourceX introductions, use the program's own rule: first valid introduction, verified application, one partner per company. The stage-to-cash reconciliation carries the report past signature into payment, and the address-book waterfall covers the step before introduction.
Where this report could still mislead
- Small cohorts. With 17 companies, one outcome moves every line. Printing counts rather than rates is the defense, not a cure.
- Long resolution. The credit window closes 90 days after the recorded introduction, though duplicate review can run longer, and payment depends on acceptance, cleared funds and the supplier being paid. A cohort can look idle for months and still pay later.
- Thin visibility. Partners see generic stages. A company stuck "in review" may be waiting on its own CEO, and the report cannot tell.
- Changing terms. The terms carry a version history; record the version in force at each cutoff, because a rule change can move a company between rows.
- Counts are not founder value. A clean pipeline can still be a program founders find tiresome. Measure that separately, as separate founder and revenue scorecards describe.
Methods note. The traps, stages and checks are built from the program terms (version 2026-11-04), the rewards page and the unsigned Channel Partner Agreement sample, all checked 2026-11-04. The waterfall and roll-forward are invented to show the arithmetic. Nothing here measures how many introduced companies are credited, accepted or paid across the program; SourceX has not published those figures.
Which companies belong in the denominator at all
Only companies someone actually introduced, with the CEO's agreement, belong in the count. A company is a plausible introduction when it has 50+ full-time employees at peak (contractors excluded), years of records it created itself, the rights to license them, and a CEO who decides whether the company discusses licensing and signs for 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 70% of the data is in English. Partners screen at metadata level and never handle records or credentials. Funding status plays no part, so a report sorted by valuation measures the wrong thing. Related reporting guides sit in the guides hub.
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.
Once your tracker counts each company once and a CEO has agreed to the conversation, get a partner link, run the company through the who qualifies criteria and note the program terms version on the report.
Reporting the fund's own referral channel is separate from judging SourceX as a company to invest in; readers doing the second can request investor materials through the SourceX investor overview, where SourceX says it shares company and commercial metrics with investors directly.
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.