The scarce input is platform attention, and a blended score spends it badly. It rewards the program that produces visible pipeline, even when founders found it a chore, and it punishes a useful program for the months in which no money could possibly have arrived. Two cards fix both errors.
Two outcomes with different clocks, owners and directions
| Founder value card | Referral income card | |
|---|---|---|
| What it measures | Whether founders found the program worth their time, and what it cost in hours | Introductions, stage changes and rewards by status |
| Who judges it | Founders, asked by someone with no stake in any reward | A ledger built from dashboard statuses and payment records |
| When it moves | Within a month of each conversation | Months later, if at all |
| Whose it is | The founders' and the platform team's | Whichever entity registered as the partner, treated as the fund's documents require |
| Can it turn negative? | Yes: wasted hours, felt pressure | It can be reduced or reversed when the supplier payout is |
| Usual failure | Politeness inflates ratings | Estimates are read as income |
The clock on the right is long by design. SourceX's terms stack five conditions in front of any payable reward: the Acquisition Partner accepts the dataset, its payment reaches SourceX in cleared funds, SourceX pays the company, finance approves, and the partner's tax and payment details are on file; the deadline then lands 60 calendar days after the Acquisition Partner's final written acceptance or completed handoff, whichever comes later. Until the company is paid, any reward shown is an estimate, and the terms call portal and calculator estimates projections, not promises. The partner dashboard reflects this with four statuses, Estimated, Eligible, Approved and Paid, and shows Paid only once an actual payout has been recorded.
The benefit on one side, the hours on the other
Define the benefit narrowly so it can be measured. For the CEO of a company that may fit, the benefit is finding out whether its own operating records could be licensed for AI training and evaluation, plus a conversation with SourceX if the CEO wants one, with the fund never handling a record. A company fits when its full-time staff reached 50 or more at peak (contractors excluded), it holds years of records it created itself and the rights to license them, its location passes the who-qualifies rule, and its CEO authorizes the discussion. The platform team never touches the records; how-it-works says a partner need not export, upload or describe confidential records, and discovery stays at metadata level. Whether a company's records match current demand is unknown until SourceX reviews it.
The cost has two parts. Platform hours go into screening, briefing whoever owns the CEO relationship, the CEO conversation and follow-up. Founder hours go into that conversation and any application. Both belong on the founder card, because both are paid in the same currency founders care about.
The dual-outcome scorecard
Card A, founder value and effort. Reviewed monthly. Thresholds are set before the first conversation and are Illustrative here; choose your own and publish them to the team.
| Measure | Definition | Collected by | Threshold (Illustrative) |
|---|---|---|---|
| Opt-in rate | CEOs who asked for a conversation, divided by CEOs offered one | Platform team | 1 in 4 or better |
| Worth the time | Share answering yes to "Was this worth the time it took?" | Someone outside the deal team | 70% or more |
| Acted on it | Founders who did something with what they learned, including deciding not to proceed | Same person, 30 days later | Recorded, no target |
| Pressure reports | Any founder who says they felt pushed, were asked again after declining, or were asked in a board setting | Same person, plus an anonymous route | Zero; one report pauses the program |
| Hours per founder helped | Platform hours divided by founders who rated it worth the time | Platform time log | 12 or fewer |
| Founder hours | The founder's own estimate | The follow-up question | Recorded, no target |
Card B, the referral ledger. Reviewed quarterly. It holds counts and statuses, never forecasts.
| Line | Definition | Source | Rule |
|---|---|---|---|
| Introductions recorded | Written introductions sent with the CEO's agreement | Your log and dashboard | Count only |
| Verified applications | Companies that applied | Dashboard stage | Count only |
| In SourceX review | Companies under review | Dashboard stage | Count only |
| Deals closed | Companies whose licensing deal reached closing | The company or the dashboard | Count only, no value |
| Rewards by status | Estimated, Eligible, Approved, Paid | Dashboard | Only Paid is income |
| Receiving entity and treatment | Who registered as the partner; whether income is disclosed, offset or declined | Fund counsel | Settled before the first introduction |
Illustrative, to show how Card B fills: if one company's licensing eventually produced an Actual Supplier Payout of USD 250,000, the reward on that first qualifying transaction would be USD 62,500, and at USD 400,000 it would be USD 70,000, with no cap for that company; a further qualifying transaction first agreed within 12 months of the first Closing would pay 15% of its own payout. Either figure enters the income line only when its status reads Paid.
Reading the two cards together
| Founder card | Income card | Decision | Why |
|---|---|---|---|
| Meets every threshold | Nothing paid | Continue | The base case: the program earns its hours on founder value alone |
| Meets every threshold | Rewards paid | Continue; disclose and treat the income as the fund's documents require | Income is a byproduct, not the reason |
| Below threshold | Nothing paid | Stop | No founder value and no reason to keep the hours |
| Below threshold | Pipeline or paid rewards | Stop anyway, and check whether pursuing rewards caused the weak scores | The case a blended score gets wrong |
| Any pressure report | Anything | Pause and investigate | Founder trust is worth more than any reward |
| Receiving entity unsettled | Anything | Pause new introductions | Income nobody can account for is a liability |
That is the whole rule: income can stop a program, through a pressure signal or an unresolved conflict, but it never keeps alive a program the founder card has failed.
Six months of a zero-income pilot (Illustrative)
Illustrative: an invented fund with 30 portfolio companies, nine of which pass every fit fact the platform team can confirm. No real pilot's ratings, hours or stage counts have been published; the table shows the method.
| Month | CEOs offered | Opted in (total) | Rated worth the time | Pressure reports | Platform hours | Introduced / applied / in review | Rewards paid |
|---|---|---|---|---|---|---|---|
| 1 | 9 | 3 | 3 of 3 | 0 | 16 | 0 / 0 / 0 | $0 |
| 2 | 9 | 5 | 4 of 5 | 0 | 12 | 2 / 1 / 0 | $0 |
| 3 | 9 | 6 | 5 of 6 | 0 | 9 | 3 / 2 / 1 | $0 |
| 4 | 9 | 6 | 5 of 6 | 0 | 6 | 3 / 3 / 2 | $0 |
| 5 | 9 | 6 | 5 of 6 | 0 | 5 | 3 / 3 / 2 | $0 |
| 6 | 9 | 6 | 5 of 6 | 0 | 5 | 3 / 3 / 2 | $0 |
At month six the founder card passes on every line: six of nine CEOs opted in, five of six rated it worth the time, nobody reported pressure, and 53 platform hours across five founders helped come to 10.6 hours each, under the limit of 12. The income card shows two companies in review and nothing paid, which is what the payment conditions predict at this point, not a sign of failure. The decision is to continue for two more quarters, and it is made without looking at Card B.
When the income card should stop the program (Illustrative)
A second invented fund runs the same design. By month four its ledger looks busy: four introductions, three companies in review, one deal closed, no reward paid yet. Its founder card does not: seven of twelve CEOs opted in, only three of seven rated the program worth the time, and one founder reported that a deal partner raised data licensing at two board meetings in a row after the founder had said no.
A blended score that credited pipeline would rank this the stronger of the two programs. The separate cards say pause at once because of the pressure report, and stop at the next review because usefulness has been below threshold for two months. Stopping does not strand the companies already in process: the terms let SourceX communicate directly with a referred company without affecting valid attribution, and stopping new asks does not require ending the partner agreement. The review that follows should ask one question first: was anyone's compensation or standing tied to the ledger?
Decide who receives any reward before the first introduction
ILPA's Principles 3.0 say all fees charged to portfolio companies should be fully offset against the management fee and accrue to the benefit of the fund, that fees outside the offset should be disclosed to investors, and that fees paid by portfolio companies to operating partners affiliated with the general partner should be offset. A SourceX reward is paid by SourceX and never deducted from the company's payout, so it is not a fee charged to the portfolio company in that sense. LPs who follow ILPA's reasoning will still ask where income connected to portfolio companies goes.
The adviser's side matters too. The SEC's 2019 interpretation of the standard of conduct says investment advisers owe each of their clients a fiduciary duty under the Advisers Act, viewed in light of the agreed scope of the relationship, and SourceX's own home page tells licensed professionals to check their rules on referral fees. Settle three things with counsel and write them on Card B: which entity registers as the partner, whether any reward is disclosed, offset or declined, and who in the fund is told. SEC scrutiny of private fund fees and referral rewards and operating partner referral fees go further.
Triggers agreed before the first founder conversation
| Trigger | Decision | Card that decides |
|---|---|---|
| Every Card A threshold met at two consecutive reviews | Offer the program to the next group of companies | A |
| Any pressure report | Pause new asks until it is reviewed | A |
| Receiving entity or treatment unsettled | Pause new introductions | B |
| Worth-the-time share below threshold at two consecutive reviews | Stop | A |
| Estimated rewards reported as income | Correct the ledger; the program itself is unchanged | B |
| Paid rewards concentrated in companies whose founders rated the program poorly | Stop and review incentives | A and B |
When the bottleneck is how many companies the team can prepare at once, prioritizing introductions when review capacity is limited supplies the queue. Whether to start a program at all is weighed in the opportunity cost of another founder program, and the first cohort's design is in choosing a ten-company pilot.
Where separate scorecards still mislead
- Small samples. Six founders produce six ratings. Treat a difference of one rating as noise.
- Polite founders. Ratings inflate when the person asking also controls the next round. Use someone outside the deal team and keep the anonymous route open.
- Lag runs both ways. A stopped program can still produce an earned reward later, because earned rewards remain payable when their conditions are met; a continued one may never produce one.
- Criteria move. SourceX applies criteria program by program and says a company that does not fit one may fit another later, so the pool of companies a founder card can reach changes over time.
- Demand is decided in review. A founder who found the conversation useful tells you nothing about whether SourceX will accept the company's records.
- Concentration. There is no cap, and the USD 50,000 floor applies once per company, so one company's outcome can dominate the ledger, which is one more reason not to steer by it.
Method note. The two cards, thresholds and decision rules are this page's own framework, applied to Illustrative funds. Program facts come from SourceX's terms (Version 2026-10-22.2, since replaced by version 2026-11-04), its how-it-works page and its screening page; governance context comes from ILPA Principles 3.0 (2019) and the SEC's 2019 interpretation; all were checked on 26 October 2026. Not measured: any real pilot's founder ratings, hours or stage transitions, and SourceX's acceptance rates or payouts, none of which is published. The broader case for judging introductions on quality rather than count is in measuring referral quality instead of volume, and the fund-side overview is for venture platform teams. Related guides are collected 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 someone with no stake in any reward is ready to collect the founder ratings, register as a partner, screen the first companies against who qualifies and read the payment conditions in the program terms before the first founder conversation.
How a platform scores its referral program is one decision and whether to invest in SourceX is quite another; for the second, the SourceX investor overview handles requests for investor materials.
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.