This page assumes the number has already earned a place in your draft. Whether a metric belongs in an introduction at all, and how to define it, is the job of the growth metric definition review. Here the questions are narrower: is the claim true, is it current, and is it yours to repeat?
Where a repeated growth number picks up its errors
A founder's growth claim rarely reaches SourceX in the founder's words; each hop strips something off.
| Hop | What usually gets lost | Ledger column that keeps it |
|---|---|---|
| Founder to call notes or a board update | The exact wording ("about", "nearly", "including our partners") | Claim, verbatim |
| Board update to the platform team's tracker | The date the founder said it, and the period it covers | As-of date and period |
| Tracker to a portfolio list or newsletter | The terms the number arrived under | Permission |
| List to an introduction email | Whether anything outside the founder's own word backs it | Corroboration |
| Email to the company's application | Who is speaking: the partner or the company | Treatment |
The signal itself is an interested party's statement about a change between two dates: useful for discovery, weak as evidence. The source and date matter more than the number.
The claim-source-date evidence ledger
Keep one row per claim, not per company, because each claim can end in a different treatment.
| Column | What you write | Why it matters |
|---|---|---|
| 1. Claim | The founder's words in quotation marks | Paraphrase drifts: "tripled" becomes "300% growth" by the third hop |
| 2. Speaker and channel | Who said it, and where: a call, an investor update, a press release | Decides whether you may repeat it |
| 3. As-of date and period | When it was said, and the two dates the change runs between | Growth claims go stale quickly |
| 4. Source class | A: a public record; B: company material the CEO cleared for reuse; C: material received in confidence; D: an oral or informal remark | Sets the default treatment |
| 5. Corroboration | A second record that agrees, disagrees or is silent, with its own date | A claim with nothing behind it stays the founder's claim |
| 6. Fit fact touched | Peak full-time headcount, years of records, systems, rights, an authorized CEO, or none | A claim that touches no fit fact need not travel |
| 7. Permission | Cleared to share, share only with consent, or internal only | Keeps confidential numbers out of emails |
| 8. Treatment | State, attribute, CEO states it, or hold | The only column the introduction sees |
| 9. Recheck date | When to ask again before reusing it | Prevents last year's number from resurfacing |
Where corroboration can come from. For a US company that sponsors a retirement or welfare plan, the plan's annual Form 5500 report is one dated, independent record. The Department of Labor's instructions for the Form 5500-SF put total participants at the beginning of the plan year on line 5a and at the end on line 5b, and the form carries a separate active-participant count. The Department's Form 5500 search is open to the public without registration and can be searched by plan sponsor or EIN. Participants are not employees: the form counts active participants separately from the total, so the total can include people who no longer work there, and the plan's own rules decide who counts at all. Use it to check direction and scale, never to replace the company's own count. Other corroboration includes the company's dated press releases and, internally only, what the same founder said in earlier updates.
Why permission gets its own column. Venture funds receive much of their growth data in confidence. In NVCA's model management rights letter, confidential information an investor receives under its management rights is subject to the confidentiality provisions of the Investors' Rights Agreement, and the company may withhold board materials to protect highly confidential proprietary information. Your fund's own documents govern, but the default is plain: a number from an investor update is class C until the CEO says otherwise.
Four treatments, and when each applies
| Treatment | Use it when | How the sentence reads |
|---|---|---|
| State with source and date | A class A record supports exactly what you say | "Its plan filing reported about 230 participants at the end of 2025." |
| Attribute to the CEO | The CEO cleared the number for reuse and nothing contradicts it | "The CEO says the company reached 150 full-time employees in 2024." |
| CEO states it | The number came in confidence, or it sits near a threshold that matters | "The CEO will give headcount and record history in the application." |
| Hold | The claim is oral, contradicted, stale or touches no fit fact | Nothing |
Three rules keep this honest. Repeating a claim never upgrades its class: a dinner remark does not become fact because your newsletter printed it. A contradiction means hold and ask, not average. And the date travels with the number.
A tripling software company and a shrinking distributor in one ledger (Illustrative)
Illustrative. Generic companies, invented claims and filings, no outcome implied.
| Claim | Speaker, channel, date | Class | Corroboration | Treatment |
|---|---|---|---|---|
| "We tripled headcount to 180 in 18 months." | Founder-CEO of a 5-year-old vertical software firm, August 2026 investor update | C | Plan filing: 64 participants at the start of 2025, 121 at the end; direction agrees, scale is lower | CEO states it |
| "That 180 includes our offshore contractor team of 45." | Same CEO, reply to a follow-up email, September 2026 | B | None needed; it narrows the first claim | Internal note; contractors do not count toward the size test |
| "We were 210 people before we sold the West Coast division." | CEO of a 23-year-old industrial parts distributor, remark at an LP dinner, June 2026 | D | Company press release dates the divestiture to 2024 but gives no headcount; a 2019 plan filing shows 198 participants at year end | Attribute only with consent, otherwise CEO states it |
| "Revenue is up 40% since the reset." | Same distributor CEO, same dinner | D | None | Hold; it touches no fit fact |
Now run both companies through the same suitability questions:
| Question | Software firm | Distributor |
|---|---|---|
| Peak full-time headcount, contractors out | About 135, if the CEO confirms | About 210 in 2019, by the CEO's account |
| Years of records the company itself holds | About three in the current CRM; earlier work in spreadsheets | 22 years in one ERP |
| Who owns the records | Much product data sits in customer tenants | The company's own, except what left with the divested division |
| Who authorizes | The founder-CEO | The CEO, with the family office owner on the board |
| What the growth claim adds | Nothing beyond a headcount the CEO must state anyway | Nothing; the decline does not erase the peak |
The fast grower's headline claim never reaches SourceX, and the shrinking firm turns out to hold the stronger facts. Neither result needed the growth number.
Reading a growth claim for the records it implies
Even a verified claim says little about records until you ask which system captured the work during the period it describes. Tripling in 18 months usually means most of the company's history is recent, and some of it may live in tools adopted mid-growth. Shrinking keeps the old history but can send part of it out of the door with a divested unit, which turns a size question into an ownership question; the founder ownership guide shows how to hand that to SourceX's rights review. The headcount growth brief reads organic, acquired and shrinking headcount the same way.
The size test is read at peak, so a smaller company today can still meet it. Write the peak and its year in column 6, not the current count.
Spend verification time where a claim changes a decision
Platform teams have more claims than hours, so work them in order:
- No conversation without the CEO. If the CEO has not said they want to talk, stop: verification cannot substitute for consent, and SourceX's screening page lists introductions made without the decision maker's knowledge among the cases that usually do not fit.
- Claims near a threshold. "About 55 people" with contractors unexplained is worth a corroboration check and a definition question; "about 600 people" is not.
- Claims that are the only evidence of history. Ask which system and which years, at metadata level, and never for files or samples.
- Everything else. Log it, set a recheck date and move on.
The relationship owner, not the platform team, asks the CEO. One partner is credited per company, so agree inside the fund who submits before anyone repeats anything outside it, and make sure that person has signed the SourceX Channel Partner Agreement.
Where the ledger misleads
- Proxies are not counts. Plan participants, press-release figures and job postings each measure something else, on their own dates. A record that agrees on direction can still be far off on level.
- Precise claims can be precisely wrong. A number with a decimal point is not better sourced than a round one.
- It can turn into diligence. SourceX's review tests size, records, rights and authority; a partner who tries to settle all of that first spends goodwill and duplicates work.
- Demand is untested. A fully verified company can still be declined, because SourceX checks relevance to its current data needs in review.
Methods note. The ledger's columns and treatments are this page's method. The fit facts come from SourceX's program terms and screening page; the corroboration example from the Department of Labor's Form 5500-SF instructions and its public search help; the confidentiality default from NVCA's model management rights letter. All were checked on 2026-10-26. The ledger does not measure demand, record quality or any payout, and SourceX has published no data on how often partner-reported growth claims prove accurate.
A measurable next step. Take the last ten growth numbers your team repeated outside the fund and fill columns 1 to 9 for each. Count how many had an as-of date, any corroboration and documented permission. If fewer than half had all three, add the ledger to the step before any introduction leaves the building.
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 a ledger row shows a CEO who wants the conversation and facts you can repeat honestly, register for a partner link, test the company against who qualifies or the company fit checker, and read the program terms before the introduction goes out. For how the rest of the program works inside a fund, see the venture platform overview.
Checking a founder's claims before an introduction is referral work; evaluating SourceX as a possible investment is a different exercise, and the SourceX investor overview is where to request the investor materials it calls for.
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.