# Compare referral candidates by growth stage | SourceX

Canonical: https://refer.sourcex.si/guides/how-to-compare-referral-candidates-across-different-growth-stages
Last updated: 2026-11-04

Compare candidates on one set of questions and change only the evidence you accept at each stage. A stage label says how a company is financed or how fast it is moving. The program asks other things: 50+ full-time employees at peak, years of records the company created, rights to license them, a CEO who wants the conversation and fit with a current program. Score every company against its own peak and its own record span, give growth rate no weight, and a contracting 30-year-old firm can be ranked fairly beside a fast-scaling six-year-old.

## What a stage label records, and what the program asks instead

Stage vocabulary comes from financing. Andreessen Horowitz, for example, describes itself as investing in seed to venture to growth-stage technology companies; those words name the round a company raised, not the records it kept. Operators add their own labels, such as scaling, mature or in decline, which describe momentum. Neither tells you whether a company fits a data licensing introduction.

Even "age" depends on who measures it. The U.S. Census Bureau's Business Dynamics Statistics assign a firm the age of its oldest establishment at birth and publish firm ages in bands that run up to 26 years or more. A screen for data introductions needs a third definition: the span of records the company created and can still read. That span can be shorter than the company's age, after a lost migration, or longer than its legal entity, after a restructuring.

This page sorts companies into five stages, each defined by headcount against the company's own peak:

| Stage | Definition used here | What the label tempts you to assume |
|---|---|---|
| Pre-scale | Has never reached 50 full-time employees | That a strong round makes it a candidate |
| Scaling | Headcount rising and at or near its peak today | That momentum means deep records |
| Steady | Headcount roughly flat for several years | That a quiet company has little worth describing |
| Contracting | Headcount well below an earlier peak | That a shrinking company has fallen out of fit |
| Exited or winding down | Acquired into a group, merged or closing | That a company that changed hands or is closing is out of scope |

The last two assumptions are the most expensive. SourceX's referral home page says operating, acquired or wound down companies can all qualify, while noting that not every company does.

## The stage-normalized candidate scorecard

Five questions, each scored 0, 1 or 2, plus a timing flag that is recorded but not added in. The point scale is this page's reading aid, not a SourceX criterion or threshold.

| Question | 2 points | 1 point | 0 points |
|---|---|---|---|
| Peak full-time headcount, contractors excluded | 50+ at peak, with a dated source such as payroll or a filing | Probably 50+ at peak, not yet documented | Never reached 50 |
| Record span | Several years of company-created records, readable end to end | Several years, with known gaps | About two years or less, or unknown |
| Continuity | The span survived every migration and ownership change | One known break | History before a migration or sale is gone |
| Rights | Material the company created itself | Company-created material mixed with client-owned material | Mostly client-owned material or personal consumer data |
| CEO path | The CEO knows and wants a conversation, through a confirmed relationship owner | A relationship owner is identified but has not asked | No path to the CEO |
| Timing flag, recorded but not added | A documented event with a date: a deletion, a system switch-off, a wind-down deadline | An expected event with no date yet | No known event |

Four normalizing rules make companies at different stages comparable:

1. **Peak, never today.** Score the highest full-time headcount the company reached, from a source dated to that year. For a staffing firm, only its own full-time employees count; placed workers on assignment at clients do not, so describe both numbers and introduce the firm anyway when its records are strong.
2. **Records, not incorporation.** Measure the span from the earliest company-created record still readable, not from the founding date or the current entity's formation.
3. **No credit for momentum.** Growth rate, funding stage and press coverage score zero.
4. **A zero is a routing signal.** A zero on rights usually ends the row. A zero on peak headcount sends it to a separate conversation, because SourceX's screening page says some programs set higher thresholds and smaller specialists can still be considered.

**Methods note.** The questions restate the published fit facts (SourceX terms and who-qualifies page, checked 2026-10-26) in units that do not depend on stage. The point scale, stage definitions and rules are this page's method. The scorecard does not measure demand, record quality or value, and nobody has tested whether its scores predict acceptance; treat it as a way to order your own asks.

## Evidence that counts at each of five stages

The questions stay fixed. What changes is where the answer comes from and which mistake is most likely.

| Stage | Headcount evidence to accept | Record span to look for | Usual trap | Who speaks for the company |
|---|---|---|---|---|
| Pre-scale | Today's payroll | Founding to now | Treating a celebrated round as a fit fact | The founder-CEO |
| Scaling | Today's payroll, contractors and outsourced teams excluded | Short, often split by a recent system change | Counting contractors toward the 50 | The CEO, often stretched by hiring or a raise |
| Steady | Payroll from any year in the flat period | Long; ask about retired systems | Skipping it because nothing is in the news | Often an owner-CEO |
| Contracting | Payroll or filings from the peak year | Long, densest around the peak | Scoring today's smaller headcount | The CEO, sometimes with lenders or advisers involved |
| Exited or winding down | Peak before the sale or closure | Complete and closed, sometimes scheduled for destruction | Missing who controls the records now | The acquirer's CEO, or whoever controls a wind-down; where a court or trustee is in control, the CEO alone may not be enough |

An adviser who reviews the books monthly often holds the dated peak-year evidence a platform team lacks, as the [fractional CFO guide for contractors](/guides/fractional-cfo-for-contractors) shows. For the end of the cycle, see [winding down a company](/guides/how-to-wind-down-a-company) and [private equity portfolio wind-downs](/guides/private-equity-portfolio-company-wind-down).

## A scaling firm and a shrinking firm, scored side by side (Illustrative)

Illustrative: both firms are invented, and their scores show the method, not a prediction.

**Firm S** is a six-year-old freight audit and payment company that grew from 40 to 170 full-time employees in three years after a Series B. Its audit exception records have lived in one platform since its third year; the first two years were kept in spreadsheets on a shared drive, and some are gone. Its audits run on clients' invoices, so the exception notes are its own while the invoices are not. A board member knows the CEO but has not raised the topic.

**Firm C** is a 31-year-old premium audit firm serving commercial insurers. It peaked at 410 full-time employees in 2015 and has 150 today. Twenty years of audit worksheets and QA reviews sit in a document management system, with older files on an archived server its IT lead says is still readable. Reports go to insurer clients; worksheets and QA reviews stay internal. Its second-generation owner-CEO, introduced by one of the fund's LPs, has said yes to a call. The archived server is due to be retired next June.

| Question | Firm S | Firm C |
|---|---|---|
| Peak full-time headcount | 2: 170, its peak, on today's payroll | 2: 410 in 2015, from that year's payroll |
| Record span | 1: six years, first two partly lost | 2: about twenty years, readable |
| Continuity | 1: one break, spreadsheets to platform | 2: unbroken so far |
| Rights | 1: own notes mixed with client invoices | 1: internal worksheets beside client reports |
| CEO path | 1: owner identified, not yet asked | 2: CEO has agreed to a call |
| Total | 6 of 10 | 9 of 10 |
| Timing flag | None | Archived server retired next June |

Scored on today's size and growth, Firm S wins easily. Normalized, Firm C leads on every question but rights. Neither result predicts SourceX's decision; it only orders your conversations.

## Does the stage change the shape of the history?

It does, and the shape matters more than the stage name. A scaling company usually has a short, dense, recent history, often split by the system it outgrew. A steady company has a long, even history with retired systems to ask about. A contracting company holds its densest records around its peak, which may be years back. A company that has been sold or is closing has a complete, closed history and, often, a schedule for destroying it.

The growth signal that put a company on your list rarely says anything about that shape: a round or a hiring spree dates an event, not the records. Ask the relationship owner three metadata questions instead: where the earliest readable records start, which system replaced which and when, and whether any archive has a retirement date. Relevance to current programs is SourceX's call in review. The live [growth-versus-suitability guide](/guides/ai-data-companies-growth) covers the same caution for fast growers, and the [age-and-ownership worksheet](/guides/how-to-find-mature-businesses-in-an-early-stage-fund-s-network) helps find steady and contracting companies beside an early-stage portfolio, and the [guides hub](/guides) collects the other screening methods.

## Order the shortlist once control and capacity are in

Use four steps, in order:

1. Drop rows with a zero on rights, and route zeros on peak headcount to a separate exception conversation.
2. Sort the rest by total score.
3. Let a timing flag lift a row above a higher score only when its CEO path scores 2. A deadline with no path is a note for the relationship owner, not a reason to contact the company.
4. Check capacity with the relationship owner before asking: a CEO closing a round or running a restructuring may want the question later.

| Rank | Company (Illustrative) | Stage | Score | Timing flag | Next step |
|---|---|---|---|---|---|
| 1 | Firm C, premium audit firm | Contracting | 9 | Archived server retired next June | LP raises it at the agreed call |
| 2 | Firm S, freight audit company | Scaling | 6 | None | Board member asks after the hiring push |
| 3 | 22-year-old printing firm in a court-supervised wind-down | Winding down | 6 | Records scheduled for destruction | Hold until the person in control is identified; the CEO alone may not be enough |

The fit facts behind every row, whatever the stage: a peak of 50 or more full-time employees, contractors aside; an operating history the company has documented; rights to license; a CEO who wants the conversation; 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. Questions stay at metadata level, and partners never request or handle records. The [company fit checker](/tools/company-fit-checker) gives a preliminary, non-binding read on any row.

## Where normalizing by stage can mislead

- **Old peaks are hard to document.** A peak from a decade ago may rest on memory. Score it 1 until a dated source appears; SourceX confirms in review.
- **Thresholds vary by program.** Some programs set higher bars than 50 full-time employees, so a 2 here does not clear every program.
- **Rights.** Client work grows with any stage, and client-owned material is usually not a fit.
- **Demand and synthetic data.** Epoch AI projected that training datasets could reach the size of the stock of public human text between 2026 and 2032, and it names synthetic data, transfer learning and data efficiency as ways around that limit. If those substitutes work broadly, demand for real business records could narrow.
- **Control in distress.** Where a court, trustee or lender controls a company, the scorecard's CEO question is necessary but not sufficient.
- **Untested weights.** Record each row's result and change the weights if they mislead you.

Whoever submits must first sign the SourceX Channel Partner Agreement (anyone who signed the older version, then titled Channel Partner Non-Circumvention and Non-Disclosure Agreement, stays covered by that signed agreement), and only one partner earns credit for a company, the partner whose valid introduction arrived first in the 90-day attribution window, with the credit landing when that company's verified application comes in.

## 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 your top row has a documented peak, a readable record span and a confirmed CEO path, [register for a partner link](/register), compare the row with [who qualifies](/who-qualifies) and read the [program terms](/terms) before the relationship owner makes the ask.

Comparing candidates across stages is a referral task, and evaluating SourceX as an investment is a different decision; investor materials are available on request through the [SourceX investor overview](https://sourcex.si/investors).

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.

## Sources

- https://refer.sourcex.si/terms
- https://refer.sourcex.si/who-qualifies
- https://refer.sourcex.si/
- https://a16z.com/about/
- https://www.census.gov/programs-surveys/bds/documentation/methodology.html
- https://epoch.ai/publications/will-we-run-out-of-data-limits-of-llm-scaling-based-on-human-generated-data
- https://sourcex.si/investors
