A pilot for networks with few eligible companies: a small-network feasibility test that counts the eligible pool before anyone is asked, sends every company through one truthful protocol and judges the pilot on what a zero-transaction result can still show

Short answer

Count the pool before you design anything. When only a handful of companies in your network pass SourceX's published fit facts, a pilot cannot compare two messages or estimate a rate. What it can show is whether the work is feasible: whether your desk count was right, whether relationship owners will ask, whether founders feel free to say no and whether the protocol holds. Run every eligible company through one truthful protocol, report each case and every zero as published counts, and decide on a date fixed in advance. No results from this design are published.

Visual briefFull-size diagram
A grid of 64 Illustrative company dots: 47 gray dots fail a known fit fact, 8 orange dots are unknown and 9 blue dots pass every fact known so far; beneath, a gauge with three bands, no pilot for 0 to 2, census pilot for 3 to 15 and comparison for 16 or more, has a marker at 9 in the census band.
Illustrative network of 64 companies. The bands are design settings written before the count, not statistical rules; no company or outcome is real.

Funding status never establishes eligibility, and in an early-stage network that cuts both ways. Most of the companies you know best may fail the size test today, while the few that pass may sit outside the portfolio altogether.

Why a short eligible list changes what a pilot can learn

A comparison needs two groups large enough to differ for a reason other than chance, and nine companies cannot supply that however you split them. NCCIH's guidance on pilot studies, paraphrased from its own page, says pilot work exists to test whether an approach is feasible and acceptable before a larger study, not to test whether it works, and it lists estimating preliminary efficacy or an effect size among the common misuses of small pilots.

A referral pilot in a thin network inherits that limit. It can show how the work goes, what each ask costs and where it breaks, but not how many introductions the network "converts", and SourceX publishes no conversion rates against which to check one.

TermMeaning on this page
FrameEvery company for which someone in your network can name the person who would ask its CEO
Eligible poolCompanies in the frame that pass every published fit fact you can establish before asking anyone
Census pilotA pilot in which every company in the pool is handled once, under one protocol, and reported as its own case
Feasibility questionA question about the process that a zero-transaction result can still answer

Step zero: the desk count that decides whether to pilot at all

The small-network feasibility test starts at a desk, with no outreach. For each company in the frame, mark each fit fact pass, unknown or fail from public sources and from what the relationship owner already knows. Files, samples, logins and anything copied from board material stay out of the count.

Fit factWhat you can establish before askingDefault mark
Full-time headcount of 50 or more at its peak, contractors left outPublic team pages and the relationship owner's knowledge; a team page that mixes contractors with staff is a reason to mark unknownUnknown until confirmed
Years of the company's own recordsFounding year from a registry or the company's site; when its own systems started is rarely visibleUnknown until confirmed
Rights to license those recordsAlmost never knowable from outside; settled only in SourceX's reviewUnknown
A CEO who would want the conversationThe relationship owner's judgment, recorded as suchThe owner's view
Location and language of the recordsWhere the company operates and the language it works in, read against the published sentence belowPass or fail

On geography, SourceX's who-qualifies page gives one sentence: 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. Add no country rule of your own.

Rights stay unknown for every company, so the pool means "passes everything you can know so far". Unknown rows on the other facts are questions for their relationship owners, and they join the pool only once answered. Then read the pool size against three bands you write down before counting.

Pool after the desk countWhat to runThe reasoning
0 to 2No pilot; a dated revisit list naming the trigger that would change each companyTwo cases describe two CEOs, not a network, and each ask still spends a relationship
3 to 15A census pilot: every company, one protocol, a case-by-case reportEnough cases to see whether the process works; far too few to split
16 or moreA comparison may become thinkable, on outcomes frequent enough to count, as in comparing two outreach anglesSplitting only makes sense when each half still holds enough companies to see anything at all

The cut points are design settings, not statistics; move them if you like, but before the count.

One protocol for every company, and no variant to test

In a census pilot every CEO hears the same truthful core, delivered by the person who holds the relationship: what SourceX does, that the company's CEO decides, whether anyone in your network could be paid if a license follows and who would pay it, that the company applies itself, and that a single no ends the subject.

The variant a small network is most tempted by is never a variant. A warmer version that says the fund recommends SourceX, or that the company "qualifies", is a breach of the protocol, because only SourceX's review decides fit and the fund endorses nothing. If you want to study wording, wait until the pool supports a comparison.

You may record who made each ask (a board member, an LP, a portfolio founder who supplies the company) without testing it. With nine rows, a pattern by asker is a hunch, not a finding.

The small-network feasibility test, question by question

Six questions make up the test. Each threshold below is an example setting; write your own into the pilot plan, with its date, before the first ask.

Feasibility questionCountDenominatorFeasible if (example setting)A miss means
1. Did the desk count hold?Companies whose fit facts, as their CEO states them, still passCEOs who stated their factsAt least two in threeDesk screening needs a better source before any wider launch
2. Will relationship owners ask?Owners who agreed to askCompanies in the poolAt least halfThe pool exists on paper only
3. How did CEOs respond?Agreed to hear, declined, no reply by day 30, each listedAsks madeNo bar; reported in fullNothing by itself; it describes the network
4. Could founders say no freely?CEOs answering yes to the freedom questionCEOs who answered itEvery oneOne no or unsure triggers the stop review
5. Did the protocol hold?Breaches: a record offered or requested, a CEO not told first, fit or endorsement impliedAsks madeZeroStop and fix before another ask
6. What did it cost?Platform and relationship-owner hours per companyCompanies askedInside the range set beforehandA budget lesson, recorded beside the rest

Question 4 is the one a venture network cannot skip. A founder asked by an investor may feel a no carries a cost. After each conversation, someone outside the deal team asks one optional question (Illustrative wording): "Did you feel free to say no to this without it affecting your relationship with the fund?" Answers are yes, no or unsure, and none is challenged.

Progress is reported and never judged: verified applications, the stage words your dashboard shows, and companies closed as duplicates because SourceX already knew them. Credit needs the company's own verified application inside the 90-day window, which no pilot can produce for anyone.

Stop rules and the decision date

  • Breach stop: any protocol breach pauses every ask until the cause is found and fixed.
  • Pressure stop: a no or unsure on the freedom question pauses asks by that asker and triggers a review of how the ask was made.
  • Owner withdrawal: a relationship owner who changes their mind ends that row, with no route around them.
  • Program change: SourceX's programs set their own criteria and can change; if the published fit facts move mid-pilot, pause and recount.
  • Exhaustion: when every company in the pool has been asked once, outreach ends. Nobody is asked twice to fill a gap.

Fix the review date at day 120 after the last ask, so every introduction has passed its 90-day window. The review picks one of three outcomes.

What the test showedDecisionWhat it never justifies
Questions 1, 2, 4 and 5 met, whatever the application countKeep the protocol as standing practice and re-run the desk count when a trigger firesAnnouncing that the network's companies qualify
Question 1 or 2 missed, protocol heldRedesign the desk count or the frame, then recountAsking the same CEOs again
Question 4 or 5 missedStop, fix the cause, and decide afresh whether to continueTreating any application that followed as a success

Nine companies, one protocol (Illustrative)

Illustrative. A seed fund's network, invented counts, no real companies or outcomes.

StageCountWhat the platform lead recorded
Frame64 companiesEach with a named relationship owner
Fail a known fact47Mostly below the size test at peak; dated, with the trigger that would change each
Unknown8Sent to their owners as questions, not asks
Eligible pool9Census band
Owners who agreed to ask6 of 9Question 2 met
CEOs asked63 agreed to hear, 2 declined, 1 no reply by day 30
CEO-stated facts still passing4 of 5One team page had counted 70 people; the CEO said 38 were full-time
Freedom question4 yes, 1 unsure, of 5 answersThe unsure followed an ask made in a board meeting week
Breaches0Question 5 met
Verified applications1Dashboard stage words copied with the date read; no transaction

The review kept the protocol with two fixes: the desk count now separates contractors, and no board member asks during a board week. The eight unknowns went back to their owners, and the single application was reported as one case, not a one-in-six rate. For companies the next count should look at, start with finding mature businesses near an early-stage portfolio and the five-ring relationship map.

What a thin-network pilot cannot tell you

  • Rates and forecasts. With a pool this small, any percentage is an anecdote with a decimal point. SourceX publishes no conversion rates, and demand for a company's records is assessed only in its review.
  • Message quality. One protocol means no comparison. A later, larger pool is where wording can be tested.
  • Privacy in the report. In a nine-company pilot, even a count can point to one founder. Keep case notes to dates, stages and the founder's own words, and share them only inside the platform team.
  • Selection. Relationship owners who agree to ask may be the warmest ones, so question 2 flatters the network if read as typical.
  • Moving criteria and rights. Each program applies its own criteria, and a company that misses one may fit another later, so a pass or fail is valid on its date only; rights stay unknown throughout.

Method note. The test uses only SourceX's published fit facts and terms, both checked 2026-10-28, and NCCIH's framing of feasibility work. It counts cases, not rates, and leaves out demand, record quality and any payout. Related designs: the ten-company selection matrix, a pilot budget with reward income at zero, the five-member association pilot and, for the review once the pilot closes, what to do when a pilot has no completed transactions.

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.

If your desk count lands in the census band, register as a partner, recheck each pool company against who qualifies and read the program terms before the first relationship owner asks; whoever submits must have signed the Channel Partner Agreement first, and more designs sit in the guides hub.

Testing whether your network can make good introductions is referral work; deciding whether SourceX is worth investing in is a separate question, for which the SourceX investor overview takes 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.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Company signs and is paidThe company signs, delivers its data and is paid by SourceX under its agreement.
  4. Step 4You get your rewardthe greater of $50,000 plus 5% and 15% of the actual supplier payout, with no cap. A reward becomes payable after SourceX has paid the referred company under its signed agreement and the payout conditions in the program terms are met, including finance approval and your tax and payment details.

Common questions

Should two small funds pool their eligible companies to get a bigger pilot?

Only if both teams run the identical protocol and each company is still asked by the person who holds its CEO relationship. Pooling lists does not change the program's one-partner rule: SourceX credits a single introduction per company from dated records, so agree in writing who would submit for each row before either team asks anyone.

Is a pilot that produces no application a failed pilot?

Not by itself. A small-network pilot is judged on whether the desk count held, whether relationship owners would ask, whether founders felt free to decline and whether the protocol held. Zero applications with all four answered is a clear result; a single application reached by breaking the protocol is not.

Can we share the pilot's results with our LPs?

Counts this small can identify individual founders, so report them to the platform team and keep them there unless each CEO concerned agrees. Never present them as a rate, a forecast or evidence of what the network will earn.

Sources

  1. refer.sourcex.si/terms
  2. refer.sourcex.si/who-qualifies
  3. refer.sourcex.si/how-it-works
  4. nccih.nih.gov/grants/pilot-studies-common-uses-and-misuses
  5. sourcex.si/investors

Free resources

Know a company with valuable proprietary data?

Become a referral partner, get your link and introduce the company’s CEO (in an owner-run company, usually the owner). Referral partners can join from most countries. We can't work with partners in countries or regions under U.S. sanctions or on U.S. restricted-party lists. The companies you introduce should be based in the U.S., Canada and the EU, or in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia with at least 70% of their data in English.

Refer a company →

I own a business

Explore licensing your company's data directly to SourceX for AI training and evaluation. Start a short assessment; no uploads needed.

Start an assessment