A platform team rarely runs out of names. It runs out of hours, and the hours run out quietly: a dozen half-prepared companies, relationship owners waiting to hear back, CEOs who were told "soon" in the spring. The rubric below is built to stop that from happening, with a zero-income base case throughout.
Hours are the constraint, and the introduction starts a clock
Review capacity here means the hours your team can give to screening companies and preparing introductions, plus the time of the partners, LPs or founders who own each CEO relationship. It does not mean SourceX's own review, which you do not control and whose timing SourceX does not publish. The founder benefit you are rationing is specific: a clear answer on whether the company's own records could be licensed for AI training and evaluation and, if the CEO wants it, a conversation with SourceX.
The program adds a clock that makes the order of work matter. Credit needs a verified application within 90 days of the recorded introduction, and the terms say the window does not extend because contact protection under the agreement continues. The sample agreement also requires introductions to be recorded in writing and says oral ones must be confirmed in writing before protection attaches. So an introduction is not a placeholder you send early to hold a spot. Sent before the CEO is ready, or before your team has time for the follow-up, it uses up days that cannot be recovered. The attribution window explainer covers the clock in detail.
How long a company waits once a line forms
John Little's 1961 proof in Operations Research established that, under stable conditions, the average number of items in a system equals their arrival rate multiplied by the average time each one spends there: L = λW. Sizing how many companies one person can carry is worked through in how many active referrals one operator can manage. This page uses the same identity for a narrower question: how long the next company in line will wait, and therefore which one should take the slot.
If four companies are in preparation at any time and each spends about four weeks there, the team completes about one introduction a week: λ = L / W = 4 / 4. Naming more candidates does not raise that number; it only lengthens the line in front of it. With twelve companies waiting and one leaving the line each week, a company added today waits about twelve weeks for its turn. By then its fit facts are three months old and its CEO's interest may have cooled. That is the case for a work cap and for a rule that stops intake.
The capacity-constrained queue rubric
The rubric has two parts: stages with caps, and a score that decides which waiting company takes the next open slot.
Part 1, stages and caps. The caps are Illustrative; set yours from your weekly hours.
| Stage | Entry test | Cap (Illustrative) | Leaves the stage when | Stale after |
|---|---|---|---|---|
| Pool | Anyone names the company | None | It is screened | Not tracked |
| Screened | Fit questions marked pass, unknown or fail with a date, from conversation only | None | A relationship owner agrees to ask, or a fail is recorded | 90 days, then screen again |
| Ready to ask | No known fail and a named relationship owner | 12 | The owner asks the CEO | 6 weeks without an ask |
| In preparation | The CEO wants the conversation | 4 | A written introduction is sent and the CEO knows who will apply | 4 weeks |
| Introduced, clock running | A written introduction is recorded | 6 | The company applies, or day 90 passes | Day 60 with no application date agreed |
| Closed out | Applied, declined or expired | None | Not applicable | Not applicable |
Part 2, the ranking score. Score each company in Ready to ask from 0 to 2 on five criteria. When a slot opens in preparation, the highest score takes it; ties go to the company that has waited longest.
| Criterion | 2 points | 1 point | 0 points |
|---|---|---|---|
| Path | The relationship owner has agreed to ask | Owner identified, not yet asked | No owner |
| Fit facts | All four known and passing | Some unknown, none failing | Not scored: a known fail removes the company from the queue |
| CEO readiness | The CEO has said they want the conversation and who will apply | The CEO is aware and undecided | The CEO is not yet aware |
| Dated event | A documented migration, retention change or archive deletion within six months | Documented, more than six months away | None |
| Effort | One call | A short chain of introductions | A group of entities that needs an entity map first |
| Expected reward | Not scored | Not scored | Not scored |
The four fit facts are the published ones: 50 or more full-time employees at peak with contractors excluded, years of records the company created itself, rights to license them, and a location that meets the rule on the who-qualifies page. The screening page lists five questions; the fifth, whether the CEO knows about and wants the conversation, is the readiness row here. It also says unknown is not a fail: an unknown is a question for the relationship holder. Every answer comes from a conversation, never from files, samples or logins, and whether the records match current demand stays unknown until SourceX reviews the company.
Work at each stage, priced in hours and weeks
Planning estimates, Illustrative. SourceX has published no stage conversion rates, so the table assumes every company reaches every stage, and income is $0 throughout because a reward exists only after a company's deal closes, SourceX pays the company and the remaining payment conditions are met.
| Stage of work | Who | Hours per company | Elapsed weeks | Income at this stage |
|---|---|---|---|---|
| Screen from what the team already knows | Platform analyst | 1.0 | 1 | $0 |
| Brief the relationship owner | Platform lead | 0.5 | 1 | $0 |
| Owner asks the CEO | Partner, LP or founder | 0.5 | 2 | $0 |
| CEO conversation, metadata only | Platform lead | 1.5 | 1 | $0 |
| Written introduction and agreement on who applies | Platform lead | 0.5 | Same week | $0 |
| Follow-up until the company applies | Platform lead | 1.0 | Up to 12 | $0 |
The middle four rows are the in-preparation stage: about 3 hours of work spread over four weeks. At steady state, screening one new candidate a week takes 1 hour, four companies in preparation take about 3 hours a week between them, and follow-up on introduced companies takes about another hour. Roughly 5 hours a week, comfortably inside an 8-hour budget, with room for the case that goes sideways.
Ten candidates for four slots (Illustrative)
Illustrative: generic companies with invented facts.
| Candidate (Illustrative) | Path | Fit facts | CEO readiness | Dated event | Effort | Score | Decision |
|---|---|---|---|---|---|---|---|
| 140-person claims administrator owned by an LP | 2 | 2 | 2 | 0 | 2 | 8 | Slot 1 |
| 210-person field-services group with a CRM migration set for March | 2 | 1 | 1 | 2 | 2 | 8 | Slot 2 |
| 180-person billing company with a documented archive deletion in January | 2 | 2 | 1 | 2 | 1 | 8 | Slot 3 |
| 95-person engineering firm | 1 | 2 | 1 | 0 | 2 | 6 | Slot 4 |
| 320-person distributor inside a larger group | 2 | 2 | 1 | 0 | 0 | 5 | Waits |
| Staffing firm with 70 internal full-time staff, placed workers counted separately | 1 | 1 | 1 | 0 | 2 | 5 | Waits |
| 600-person freight operator that buys from a portfolio company | 1 | 1 | 0 | 0 | 1 | 3 | Waits for an owner |
| 45-person seed-stage company | Not scored | Fails size | Not scored | Not scored | Not scored | Not scored | Removed; revisit when full-time staff pass 50 |
| 120-person agency whose records belong to clients | Not scored | Fails rights | Not scored | Not scored | Not scored | Not scored | Removed |
| 400-person software company whose CEO declined | Not scored | Not scored | Declined | Not scored | Not scored | Not scored | Closed; nobody goes around the CEO |
Three companies tie at 8 and all fit, so the slots fill in order of time waited. The distributor drops to "waits" only on effort: a group needs an entity map first, and a parent with its subsidiaries counts as one company for the reward cap, which the subsidiary reward cap guide explains. The staffing firm counts only its own full-time employees, not placed workers; describing both numbers keeps it in the queue rather than out of it.
A queue that should stop taking candidates (Illustrative)
A second invented team announces the program to its whole portfolio. A survey returns 40 names and screening moves 26 of them into Ready to ask. The team can carry three companies in preparation, each taking about six weeks, so by Little's law it completes about one introduction every two weeks. Twenty-six waiting companies at that pace means the last one waits about a year. Meanwhile two CEOs introduced in the spring have reached day 75 without an application date.
The rubric's answer is to stop intake now, not to work harder. Close the rows that have aged past their stale dates and tell each relationship owner plainly, check once with the two CEOs whose clocks are nearly out, and choose between adding hours and running a smaller program. A broad announcement created a queue the team could never serve; a ten-company pilot would have kept the line short enough to finish.
Expansion, pause and stop rules
Set these before the first ask. The thresholds are Illustrative.
| Signal | Threshold (Illustrative) | Decision |
|---|---|---|
| Expected wait in Ready to ask (companies waiting divided by weekly completions) | Over 8 weeks | Pause intake |
| An introduced company at day 60 with no application date agreed | Any | One check-in with the CEO; no second chase |
| Asks declined by relationship owners or CEOs | More than half over a quarter | Stop and revisit how companies are screened |
| Founder ratings of the conversations | Below your threshold two months running | Stop |
| Any founder reporting pressure | One report | Pause new asks |
| Expected wait under 2 weeks and founder ratings strong at two reviews | Both | Expand: raise the preparation cap by one, or add hours |
How to measure the founder side without letting income distort it is set out in why founder value and referral revenue need separate scorecards.
Where a queue rubric breaks down
- Scores are judgments. Two people will score "CEO readiness" differently. Write one sentence of evidence beside every 2.
- Little's law is an average. It describes stable, long-run flow. On ten companies it is a sanity check on the size of the line, not a forecast.
- SourceX's timing is outside the queue. Review times are not published, so never promise a CEO a date for anything SourceX does.
- Criteria move. Criteria follow each SourceX program and can change, and a screen is valid on the date it was made, which is why screened rows go stale.
- Urgency must be documented. A dated migration or deletion earns points only with evidence; a quarter-end or an enthusiastic partner does not.
- Relationship cost is invisible here. Each ask spends an LP's or founder's goodwill, and no data on how they react has been published; note it in the row.
- Demand is not scored. The rubric orders your work, not SourceX's interest; relevance to current data needs is decided in review.
Method note. The stages, caps, scores and thresholds are this page's own framework, applied to Illustrative companies. Program facts come from SourceX's terms (Version 2026-10-22.2, since replaced by version 2026-11-04), the sample Channel Partner Agreement and the screening page, all checked on 26 October 2026, and the queue arithmetic from Little's 1961 paper, checked the same day. Not measured: real platform hours, stage durations, decline rates or SourceX review times, none of which is published. The earlier step of ranking a whole network is in prioritizing a network of business owner relationships, a starting calendar is in the first-month plan for venture platform teams, and the fund-side overview is for venture platform teams. More operating guides are 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.
When the first company reaches the top of your queue with an agreed path and a ready CEO, register as a partner and sign before you submit, recheck its fit facts against who qualifies, and read the attribution window in the program terms before the introduction goes out.
Ordering a referral queue is an operating decision for the platform team; whether to invest in SourceX is a different question with a different owner, and anyone weighing it can ask for investor materials through 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.