This page is for whoever runs a bench of introducers: a firm's advisors, an MSP's account leads, the members of an alliance or a software ecosystem's partner agencies. Program facts come from SourceX's terms (Version 2026-10-22.2, since replaced by version 2026-11-04) and its who-qualifies page, both checked on 26 October 2026. The scorecard, its thresholds and the worked bench are this page's proposal with invented figures; SourceX has published no partner-level quality data.
Activation tells you who moved; quality tells you whether the move helped the company
Activation is a partner's first introduction that a company acts on. It answers "who is doing anything", and the registered-enabled-active ladder already counts it. Quality starts after that point and asks a harder question: when this partner introduces a company, does the company fit, did its CEO want the conversation, and did the partner do their small part cleanly?
Logo counts cannot answer it. A partner with nine introductions and seven avoidable misses has cost the companies, your reviewers and SourceX more time than a partner with three clean ones. The program also rewards care over volume by design: only one partner is credited per company, and a company SourceX already knew through any earlier application, relationship, lead, account or email correspondence is a duplicate that nobody is credited for.
| Term | How this scorecard uses it |
|---|---|
| Unique company | Counted once per partner however many times it was submitted; a parent and its subsidiaries share one reward cap |
| Partner-controllable miss | A company that failed a fit fact the partner could have checked at metadata level before contacting the CEO, such as peak full-time headcount, client-owned records or location |
| Outside the partner's control | Relevance to current data needs, rights found in contract review and duplicates in SourceX's own records |
| Company feedback | Four answers from the introduced CEO, collected by your team after the company applies or declines |
| Follow-through | The partner's own duties: claim the company in your log, confirm the introduction in writing, close the loop on a decline |
The fit-trust-follow-through scorecard
Each dimension scores 0 to 2 from records you can date. Trust works differently: it is checked first, and a failed gate stops the scoring. The fit facts are the published ones: a peak of 50 or more full-time employees with contractors excluded, a multi-year record the company produced on its own, the rights to license that record and an eligible 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.
Whether a company's records match current demand is checked in SourceX's review, so a decline for relevance never counts against the partner. Neither does a duplicate hidden in SourceX's records, because no partner can see them.
Read the total out of six only after the gate passes: 5 or 6, keep the partner and protect their time; 3 or 4, coach the weakest dimension; 0 to 2, cap proposals at two a month and look again next quarter.
| Dimension | Evidence (dated) | 2 points | 1 point | 0 points | Score only when |
|---|---|---|---|---|---|
| Trust gate | Company feedback question 1 and any complaint | Not scored: pass or fail | Not scored: pass or fail | Fail pauses the partner | Any reply arrives |
| Trust | Company feedback questions 2 to 4 | Every reply clean | One reply flags an overstatement, a file request or a "no" to another call | Two or more such flags | Two or more replies |
| Fit | Screen results for the last six unique companies, with the criteria version | No partner-controllable misses | One | Two or more | Three or more unique companies |
| Follow-through | Claim log, written introduction records, decline close-outs | Every case logged before contact, confirmed in writing and closed within 14 days | One lapse | Two lapses, or a submission against another partner's logged claim | Three or more unique companies |
Four questions for the CEO, sent by you rather than the partner
The partner should not collect their own feedback; a CEO who likes the person who introduced them will soften the answer, and the partner would see it. Send it from your team, once, after the company has applied or declined, with one reminder at most. AAPOR's survey guidance, in paraphrase, points the same way the questions below do: one concept per question, plain terms, no leading wording, and an awareness that people drift toward polite or agreeing answers. So every question asks what happened, not how the CEO felt about it.
Nothing in the form touches the company's records or its application. SourceX's terms say partners never see other partners' records, applicant documents or confidential buyer information, and your feedback loop should not try to.
- Before SourceX contacted you, did you know this partner was introducing your company, and had you agreed? (yes, no, not sure)
- Did the partner ask you for files, samples, exports or logins? (yes, no)
- Did the partner say your company would be accepted, or name an amount it would be paid? (yes, no)
- If this partner raised another business topic with you next year, would you take the call? (yes, maybe, no)
| Answer | What it means for the partner |
|---|---|
| Question 1 is "no" | Trust gate fails; pause new introductions and talk to the partner within a week |
| Question 1 is "not sure" | Ask the partner for the dated record of the CEO's agreement before scoring |
| Question 2 is "yes" | Trust flag; discovery stays at metadata level: what, how much, how long, which system, who owns it |
| Question 3 is "yes" | Trust flag; SourceX decides acceptance after review, and no payout exists before a completed deal |
| Question 4 is "no" | Trust flag, and the most useful sentence to read back to the partner |
A twelve-introducer bench at the half-year mark (Illustrative)
Illustrative: an invented advisory network, generic roles, no SourceX data. Twelve partners introduced 41 unique companies over six months.
The underperforming cohort. Five partners recruited at one conference sent 14 companies between them and nine were avoidable misses, mostly companies with client-owned records or too few full-time staff. Their CEOs had no complaints. Enthusiasm is not the problem; criteria are. They get one working session on their own declined rows and a two-proposal monthly cap until fit reaches 1.
The duplicate-risk case. Partner K submitted a 260-person property manager six days after partner B had logged a claim for it. One partner is credited per company, and SourceX decides credit from dated records; in your own log, B was first. The scorecard docks K's follow-through, not B's. Separately, one of A's companies had emailed SourceX a year earlier, which makes it a duplicate for everyone; that is nobody's miss.
The trust failure. Partner C's fit record is as clean as A's, yet one CEO answered question 1 "no". C is paused, not removed, until the two of you have walked through how that introduction was made.
| Partner (Illustrative) | Unique companies | Fit | Trust (replies of asked) | Follow-through | Decision |
|---|---|---|---|---|---|
| A, fractional CFO | 6 | 2 | 2 (4 of 5) | 2 | Keep; first review slot each month |
| B, M&A advisor | 4 | 1 | 2 (3 of 4) | 2 | Keep |
| C, MSP account lead | 5 | 2 | Gate failed (2 of 4) | 1 | Paused pending a conversation |
| D to H, conference cohort | 14 combined | 0 for four of five | 2 (6 of 9) | 1 | Criteria session; two proposals a month each |
| I, lender relationship manager | 3 | 2 | Not scored (1 of 3) | 2 | Keep asking for replies |
| J, former COO | 2 | Not scored | Not scored (0 of 1) | Not scored | Unproven below three companies |
| K, accountant | 4 | 1 | 2 (2 of 3) | 0 | Coach on the claim log |
| L, strategy advisor | 3 | 0 | 1 (2 of 2) | 1 | Correct how the program is described |
Review hours, support load, concentration and claim collisions
A scorecard is work, and it has to fit the hours you have. On this bench it costs about three hours a month: roughly 15 minutes per company to send, chase once and log feedback, plus a 20-minute quarterly review per partner. At 60 partners and about 200 introduced companies a half-year, the same method would need about 15 hours a month, nearly double an 8-hour budget. Past that point, score monthly only partners with three or more companies, send feedback to every company from paused or flagged partners and to one in two otherwise, and slow recruiting. The arithmetic for that decision is in when a partner network grows faster than review capacity.
Sister companies need care. Two partners may introduce different subsidiaries of one group; the reward cap is shared across the group, but the terms say nothing about whether the earlier introduction of one sibling affects credit for the other. SourceX settles that in its duplicate review, so neither partner should be told an answer in advance.
| Signal | How to compute it | Threshold (Illustrative) | What you do |
|---|---|---|---|
| Scorecard hours | Feedback and review minutes per month | Over your budget two months running | Sample feedback, score quarterly, slow recruiting |
| Support load | Partner questions per month, by partner | One partner asking a third of all questions | Answer once in writing and share it with the bench |
| Concentration | Share of fit-passing companies from one partner | Over the greater of $50,000 plus 5% and 15% | Plan for that partner's absence before relying on the number |
| Claim collisions | Submissions naming a company already in your log | Two or more in a quarter | Make the log check a required step before any CEO contact |
Evidence gates and the partner's right to correct a row
Scores drive decisions only when they rest on enough records. Below three unique companies or two feedback replies, a partner stays unscored, as J does. Every fit result carries the criteria version it was screened under: SourceX's criteria follow each program and can change, and a screen is valid on the date it was made, so a miss under last month's criteria may not be a miss today.
Show each partner their own row and nobody else's. Let them correct facts, such as a company you logged against the wrong partner or a decline caused by relevance rather than headcount, and keep the correction with its date. The scorecard governs your support, coaching and pauses only. It never decides attribution or rewards; those follow the terms, and disagreements go to SourceX with the referral details and evidence. For comparing whole partner roles rather than individuals, use the partner quality study method, and for settling which of two introductions came first, the seven evidence checks for two partners and one company.
Ways this scorecard can flatter or punish the wrong partner
This page rates introducers; rating a dataset is a different job, covered by the data quality scorecard on sourcex.si. More partner-network playbooks are in the guides hub. Each limit below is a reason to read a score as a prompt for a conversation, not a verdict.
- Small numbers swing. One miss moves a three-company partner from 2 to 1. The evidence gate limits this; it does not remove it.
- Feedback non-response is not neutral. CEOs who reply may differ from those who stay silent. Report replies against requests, as the table does.
- Politeness hides problems. Behavior questions help, but a CEO may still answer kindly. One "no" to question 4 deserves more weight than three "yes" answers.
- Fit screens are not acceptance. A company that passes your screen can still be declined in review for relevance or rights.
- The scorecard can drift into a ranking. Keep it internal and never publish scores or tie them to promises of income.
- Duplicates you cannot see. SourceX's records decide them; keep your own log clean with the canonical identity register and accept the rest. To group partners by how they behave over time rather than score them one by one, see segmenting partners by real referral behavior.
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 bench sends its next introductions, have each partner register for a partner link and sign, check their candidates against who qualifies, and read the attribution and duplicate rules in the program terms so the scorecard measures care, not luck.