This is a design for one partner or one network running its own outreach. It does not touch SourceX's form, and the partner never collects, reads or reviews any company's answers or records.
What the checklist may restate, and what it must never do
The checklist is a reprint of who qualifies in the second person, nothing more. SourceX calls that page screening guidance, and its own company fit checker a preliminary, non-binding screen. Your checklist cannot be stricter, looser or more certain than either.
| The checklist may | The checklist must never |
|---|---|
| Restate the published fit facts in plain words | Add a criterion SourceX has not published |
| Tell the CEO to keep the answers | Ask for answers, counts, files or screenshots back |
| Say that SourceX decides fit in review | Say or imply "you qualify" |
| Link the fit checker and the who-qualifies page | Score the company or give a pass mark |
| Say that smaller specialists can still be considered | Tell anyone below a threshold not to apply |
The last row matters for the experiment. SourceX's page says some programs set higher thresholds and smaller specialists can still be considered. A checklist that turns guidance into a gate could make results look better by quietly losing companies SourceX would have reviewed.
The seven self-check lines
This is the treatment, word for word, as the checklist arm receives it. Date it, and reissue it whenever who-qualifies changes.
- At its peak, did your company have 50 or more full-time employees? Contractors do not count; a staffing firm counts its own full-time staff, not the workers it places with clients.
- Has your company created its own records for several years: procedures, project histories, quality records, support workflows or similar?
- Does your company own that material, or can it license it without breaching client confidentiality, employee privacy or contracts with others?
- Is the material something other than personal consumer data, scraped data or records that belong to your clients?
- Location and language: 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. (How the English share is counted.)
- Do you, as CEO, want a conversation with SourceX about licensing? You decide; nobody else can decide this for the company.
- Do you understand that SourceX decides fit in its review, that programs differ, and that a company that does not fit one program may fit another later?
The footer, in the same message: "Keep your answers; do not send them to me. If any line is unclear, apply anyway and let SourceX's review settle it, or try the fit checker." That last instruction is deliberate. The checklist informs a choice; it does not replace the review.
Two arms, one cohort, allocation fixed before you write
| Element | Specification |
|---|---|
| Question | Does adding the seven lines change the share of companies that reach a qualified application? |
| Unit | One company whose CEO agreed in writing to hear about SourceX before any link was sent |
| Control arm | Your standard introduction message, with a link to who-qualifies |
| Checklist arm | The same message, word for word, plus the seven lines and the footer |
| Allocation | Alternate by the date order of each CEO's yes, or a coin flip logged before you draft the message; never by your sense of which company fits |
| Exclusions | Companies you already know to be duplicates, removed before allocation; nothing removed afterwards |
| Entry window | Fixed in advance, for example two quarters |
| Cutoff | 120 days after the last message, so the 90-day attribution window and some review time fit inside it |
| Primary measure | Qualified applications per allocated company, by arm |
| Analysis | By assigned arm, whatever the CEO did with the message |
Qualified application, defined. An application from an allocated company that SourceX has verified and credited to your introduction, whose stage on your dashboard at the cutoff is anything other than a closure. Copy SourceX's stage words exactly as displayed; how it works says the dashboard shows each referral's stage and its reward status. A link click, a visit or an opened email never counts, and the terms agree that a click alone decides nothing.
Measures beyond the primary count
| Measure | Why it is there | Who produces it |
|---|---|---|
| CEO declined before applying | Separates self-exclusion from silence | The CEO's own reply; record only that they declined |
| Declines that mention the checklist | The main unwanted effect: a fitting company talked out of applying | The CEO's words, only if offered |
| Days from the CEO's yes to application | Whether the checklist slows people down | Dates you already hold |
| Clarifying questions per company | Whether the lines confuse more than they help | Your inbox |
| "Felt like a test" or "thought it meant approval" | Trust and misreading | Unprompted feedback |
| Answers or records sent to you | Protocol failures; the target is zero | Your inbox |
| Duplicates | Should be similar in both arms; a gap means allocation went wrong | Your dashboard |
Record declines without the private reasons behind them; a seven-code decline taxonomy shows how.
Minimum sample, duration and stop rules
Most partners will introduce too few companies in a quarter for a confident comparison. Set the floor before the first send: compare rates only once each arm holds at least 15 allocated companies and the entry window has closed; below that, publish counts and no percentages. Partners in one network may pool their records if every one uses the same checklist version, the same allocation rule and the same cutoff. The ten-introduction experiment explains what very small counts can and cannot settle.
Stop rules.
- A company sends you its answers or any record: stop sending the checklist until the footer is rewritten, and do not keep what arrived.
- A CEO reads the checklist as an approval or a promise: stop and fix the wording before the next message.
- SourceX changes who-qualifies or the terms: stop, reissue a dated version and start a new entry window. Mixing versions ruins the comparison.
- You catch yourself steering promising companies into one arm: stop and restart with a logged coin flip.
- Never stop because one arm looks ahead. Run to the cutoff.
How the result would be read
Report every allocated company, not just the ones that applied. The CONSORT 2025 reporting guideline for randomized trials pairs a 30-item checklist with a diagram of how participants flow through the trial (paraphrase). The same discipline fits a partner's test: one table, per arm, at the cutoff.
| Count at cutoff | Control | Checklist |
|---|---|---|
| Companies allocated | Not yet collected | Not yet collected |
| CEO declined before applying | Not yet collected | Not yet collected |
| No reply by the cutoff | Not yet collected | Not yet collected |
| Applied | Not yet collected | Not yet collected |
| Verified and credited to the introduction | Not yet collected | Not yet collected |
| Found to be duplicates | Not yet collected | Not yet collected |
| Closed by SourceX | Not yet collected | Not yet collected |
| Qualified applications (primary) | Not yet collected | Not yet collected |
| Answers or records sent to the partner | Not yet collected | Not yet collected |
Decision rule, written before the first send. Agree the smallest difference worth acting on, for example ten percentage points in qualified applications, and how you will show uncertainty: a range for the difference, not a single number.
| Result | Decision |
|---|---|
| The range for the difference sits wholly above your threshold, with no rise in checklist-related declines or complaints | Adopt the checklist for all agreed CEOs |
| The range sits wholly below zero | Drop it; it costs applications |
| The range crosses zero or your threshold, or sits between them | Inconclusive: keep the control message, say so, and keep counting if you wish |
| Qualified applications equal, self-declines up, applications that clearly fail a fit fact down | The checklist filters without loss; adopting it is your judgment, stated as such |
| Any protocol failure count above zero | Fix the wording first; the comparison waits |
Cuts by sector, company size or relationship type are descriptive only, never grounds to declare a winner. A sibling test on SourceX's own headline uses the same structure at site scale in the maximum-payout headline study.
Where the checklist test could mislead
- Self-report. The CEO answers alone and you cannot see whether the answers were right. That is the price of never collecting records, and it is worth paying.
- Programs move. SourceX applies each program's own criteria. A checklist that was accurate in the spring can be wrong by autumn, which is why every version carries a date.
- Fit facts are not the whole review. Relevance to current data needs, the rights review and what a license would cover are decided only by SourceX. A checklist that predicts applications well still predicts nothing about acceptance or payment, and a lead, meeting or signature alone triggers no reward.
- Your message changes too. Adding seven lines makes the email longer. If the checklist arm wins, length is a rival explanation; a later test can compare the checklist against an equally long message without it.
- Duplicates. A company already in contact with SourceX is not credited to any later partner, whichever arm it sat in, so check before allocation, not after.
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 allocating your first CEO, register as a partner and sign the Channel Partner Agreement, copy the checklist only from the current who qualifies page and note the program terms version in your test log. Related designs are listed in the guides hub.