Platform teams usually pick a channel by temperament: the research-minded write, the community-minded host, the dealmakers call. Each choice spends a different resource, and only one of them, goodwill, is hard to get back. A small experiment shows which resource buys what in your own portfolio.
What each channel gives a founder, and what it costs the platform
The founder benefit is the same in every channel. A CEO learns whether the company's own operating records could be licensed for AI training and evaluation and, if the CEO wants it, gets an introduction to SourceX, which then reviews the company. The channels differ in what they cost and whom they spend.
| Channel | What the founder gets | Cost shape for the platform | Whose goodwill it spends | Main risk |
|---|---|---|---|---|
| Research memo | Two pages on which records can fit, how direct licensing works, the fit tests, who decides, and what the company never hands over before it signs | Mostly fixed: writing and checking once, then replies | Almost nobody's | Read as promotion, or not read at all |
| Closed founder session | Sixty minutes with 8 to 14 peers discussing their own records | Fixed hosting, plus follow-ups | The host's | Social pressure in the room; word spreads to other founders |
| Direct introduction ask | One relationship owner asks one CEO | Per company, every time | The relationship owner's | Goodwill spent one relationship at a time |
Research and sessions speak to many founders at once, so the reward has to be stated in each of them. The FTC's staff guidance on its Endorsement Guides says a material connection, including a financial relationship, should be disclosed when it could affect the weight people give an endorsement, and that one general statement on a homepage is not enough because people may never see it. The Guides are not regulations, and a memo to portfolio CEOs is not a consumer advertisement, so treat this as a working standard and ask counsel about your case. In practice, every memo and invitation says the fund or platform may receive a referral reward from SourceX if a company it introduces completes a qualifying deal, and that SourceX pays it separately, never out of the company's payout. Discussing referral rewards with a business contact has wording.
Design the channel allocation experiment
1. Build the pool. List the companies you can reach that pass the fit tests as far as you know: a peak full-time headcount of 50 or more without counting contractors, an operating history on record, rights to license, an authorized contact, 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. Funding status is not a test, so LP-owned and co-investor companies belong in the pool. Remove any company already in contact with SourceX: it is not a new company and is reviewed as a duplicate.
2. Assign by lot. J-PAL's introduction to randomized evaluations describes assigning participants at random to groups that receive different types of an intervention, so the groups are alike on average and later differences can be credited to the intervention. Two of its cautions matter here. The groups are comparable on average only with enough units, and a platform pool is small. And where spillovers are a concern, assignment can be made by group: if several companies share a co-investor or sit in one sector peer group, assign the group, not the company. Stratify by relationship owner so one partner's warm relationships do not all land in one arm.
3. Fix budgets and rules before anyone is contacted.
| Arm | What happens | Hours ceiling (Illustrative) | Primary measure | Guardrail |
|---|---|---|---|---|
| Research | The memo goes to each assigned CEO, with the disclosure; replies are answered | 20 | CEOs who ask for a conversation, per 10 platform hours | One reminder at most |
| Session | One invitation to a closed session; follow-up only for CEOs who ask | 24 | The same | No sign-up sheet in the room |
| Direct ask | The relationship owner asks each assigned CEO once | 30 | The same | A decline stands for the quarter |
Secondary measures: founder usefulness, asked two weeks after each conversation by someone outside the platform team; declines and complaints; relationship-owner hours in their own column; and company applications, recorded but not used to decide, because SourceX's review runs past the quarter.
The work in each channel by stage, with income at zero
These are Illustrative planning hours for the platform. The difference that matters is fixed against per-company cost.
| Stage | Research memo | Closed session | Direct ask |
|---|---|---|---|
| Preparing, once per arm | 8 hours to write and check the memo | 10 hours to plan and host | 1 hour for the owner brief |
| Reaching each assigned CEO | 0.25 hours | 0.5 hours for invitation and reminder | 1.5 hours to brief the owner and ask |
| Metadata-only conversation with each CEO who opts in | 1.5 hours | 1.5 hours | 1.5 hours |
| Written introduction and portal submission | 0.5 hours | 0.5 hours | 0.5 hours |
The conversation stays at metadata level in every arm: which records, roughly how many, how many years, which system and who owns it, with no files or logins, and the CEO decides whether to go further. Income is $0 in every arm, because a quarter is too short to observe it: before any reward is payable, the Acquisition Partner has to accept the data and pay SourceX in cleared funds, the supplier compensation has to be paid, finance has to approve and the partner's payment and tax details have to be on file, and the outside deadline still sits 60 calendar days past the later of completed handoff and the Acquisition Partner's final written acceptance.
Reading one quarter's results (Illustrative)
Illustrative: an invented fund with 36 companies in the pool, 12 per arm, and hours counted through submission.
| Arm | Companies assigned | Platform hours | CEOs who asked for a conversation | Opt-ins per 10 hours | Rated useful | Declines or complaints |
|---|---|---|---|---|---|---|
| Research memo | 12 | 15 | 2 | 1.3 | 2 of 2 | None |
| Closed session | 12, of whom 7 attended | 22 | 3 | 1.4 | 2 of 3 | One CEO felt pitched in the room |
| Direct ask | 12 | 29 | 5 | 1.7 | 4 of 5 | One owner declined to ask |
The direct arm also used 5.5 relationship-owner hours, half an hour for each of the 11 asks made; the other arms used none.
Read it honestly. With 12 companies per arm, one more research CEO saying yes would have moved that arm to the top, so the ranking on opt-ins is not established. Three things are: research reached 12 founders for half the hours of direct asks, the session produced the only complaint, and direct asks drew on partners' goodwill. This fund's decision for the next quarter: the memo goes to every company in the pool as the default; direct asks are kept for CEOs who reply or face a dated event such as a system migration; the session is dropped unless founders ask for one. The same design runs again with a larger pool.
Expand, pause or stop on a bar set in advance
| Result | Decision |
|---|---|
| One arm clears the bar, for example at least 1.0 informed opt-in per 10 platform hours, most opt-ins rating the conversation useful and no complaint, while core support response times hold steady | Expand that channel next quarter, within the operator's measured capacity |
| Arms within noise of each other | Keep the cheapest per founder reached as the default and rerun with a larger pool |
| An arm produces a complaint or a strained relationship | Drop that arm, whatever its opt-in rate |
| No arm clears the bar | Stop proactive outreach and handle CEOs who ask, one at a time |
How much intake one operator can carry is worked out in the stage-weighted workload calculator, and the signs that any of this is crowding out core work are in when a referral program distracts from core portfolio support.
When the experiment says stop (Illustrative)
Illustrative: an invented seed fund.
Only 14 companies pass the screen, so the arms are 5, 5 and 4. After the quarter the research arm has produced no opt-ins for 10 hours of work. Two CEOs attended the session, and one asked for a conversation: 1 opt-in for 14.5 hours, or 0.7 per 10. The direct arm produced one opt-in for 9 hours, 1.1 per 10, but an LP asked the platform not to approach that company's CEO again. Under the rules written beforehand, the direct arm is dropped for the strained relationship and the other two miss the bar. Nothing clears it, so outreach stops after 33.5 hours and $0 of income. The lesson is the pool: an experiment needs more companies than this fund can reach, so it handles CEOs who ask, one introduction at a time, and reruns when the pool doubles.
Limits of a small experiment
- Small arms. J-PAL's point about enough units bites hard at 12 per arm. Treat gaps of one or two opt-ins as ties and look for differences in cost and complaints instead.
- Spillover. Founders talk. A session attendee who mentions it to a memo-arm CEO blurs both arms; assigning by group reduces this but does not remove it.
- Season. A quarter with a fund close or year-end planning depresses every arm. Compare arms with each other, not with last year.
- The measure stops early. An opt-in is not a qualified company. Fit is decided in SourceX's review, and a no from one program is not a no from all of them, since another may suit the company later.
- Disclosure changes behavior. A clear reward statement may lower opt-ins in the research and session arms. Keep it anyway.
Method note. The channels, hours, measures and thresholds are this page's own design, applied to Illustrative funds. Program facts come from the SourceX terms (Version 2026-10-22.2, since replaced by version 2026-11-04) and the screening page, checked 26 October 2026; the assignment method follows J-PAL's introduction to randomized evaluations; the disclosure standard follows the FTC staff guidance cited below. Not measured: results from any real fund, SourceX review durations, acceptance rates and payouts, none of which is published. For session formats, see client roundtable ideas; for direct asks, warm introductions versus cold outreach and the founder email templates; and the overview for venture platform teams covers the program from the fund's side.
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 the first memo, invitation or ask goes out, register for a partner link, screen the pool against who qualifies and read the program terms, so every channel describes the program accurately; more platform guides are in the guides hub.
Choosing how to reach founders is a referral question; deciding whether SourceX is a company worth investing in is a separate one, and anyone asking that second question can request investor materials from the SourceX investor overview.
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.