This guide is for VC platform research leads who notice a portfolio or network company shipping at a striking pace and wonder whether that makes it worth an introduction. Neither funding nor launch cadence appears among SourceX's checks; the product history behind the cadence is what gets screened.
Launches, releases and deployments are three different counts
Before a cadence goes into your notes, decide which of three things it counts.
| Term | What it counts | Where a count comes from | What it says about records |
|---|---|---|---|
| Launch | A product or feature announced to customers or the market | Company announcements and changelogs | That a release cycle exists, nothing about how it was recorded |
| Release | A versioned change shipped to customers | Release notes; on GitHub, releases are based on Git tags that mark a specific point in a repository's history | That dated release records exist, if the company kept them |
| Deployment | A change put into production | DORA defines deployment frequency as the number of deployments over a period, or the time between them | Volume of change, often automated; the reasons sit elsewhere |
| Change lead time | Time from commit to production | DORA's throughput metrics | How fast work moves, not how long the history runs |
A launch count is a marketing number and a deployment count is an engineering number; neither is a record count. The records that matter for a referral conversation sit around each release: what was planned, which issues it closed, what was tested, who signed it off and what broke afterward. SourceX's page on product release coordination describes that trail as release plans, issue histories and release notes, kept in tools such as Jira, Linear and release tools.
The signal itself usually reaches a research lead in one of four ways: a public changelog (a dated company disclosure, useful for cadence), public repository releases (visible only for public code, which says little about private products), board updates (confidential, so a reason to ask the CEO and never material to pass on), or the company's own account, which is the only one that settles anything.
The product-history referral interview
Ten questions in three blocks, for the head of engineering, head of product or release manager the CEO nominates. Record each answer with its source, the person and the date.
| # | Question | Why it matters | Who usually answers | Never ask for |
|---|---|---|---|---|
| 1 | When did the company start tracking releases in a system, and which one? | Sets where the history begins | Head of engineering | Repository access |
| 2 | Has release tracking changed tools, and where does the older history live now? | A migration splits history and creates a dated event | Engineering operations | A login to the old tool |
| 3 | Roughly how many releases has the company shipped, and in which years? | Spreads the cadence across time | Release manager | Release notes in bulk |
| 4 | What does one release leave behind: plan, issues, test results, sign-off, notes? | Connected records show the path from request to result | Release manager | An example release |
| 5 | Are rollbacks, incidents and defects linked to the release that caused them? | Outcomes show how the work ended | Head of engineering | Incident reports |
| 6 | Who makes the calls: approvals, change reviews, named roles? | Judgment calls and handoffs | Head of product | Staff names |
| 7 | Who writes the code and tickets: employees, contractors or an outsourced team? | Contractors are excluded from headcount, and their work raises rights questions for review | CFO or head of engineering | Contracts |
| 8 | Do any releases live in customers' environments, or include builds that customers own? | Client-owned material is usually not a fit | CEO, with counsel if unsure | Customer names |
| 9 | What was full-time headcount at its peak, contractors excluded? | The size test | CFO | Payroll files |
| 10 | Does the CEO want the conversation, and when is the next month without a launch? | The CEO authorizes; launch weeks leave no capacity | CEO | A commitment |
Questions 4 to 6 carry the most weight. SourceX's workflow page says release records are valued because they show the full path from request to result and capture judgment calls and handoffs, and that years of history show many variations of the same task. Its process page lists historical depth and how work ended among its ten fit signals. A company that ships daily but records only the deployment has answered question 3 well and questions 4 to 6 poorly. For the counters and admin reports that help with question 3, see investigating record growth without record access.
Two release patterns under the same ten questions (Illustrative)
Illustrative. Two invented software companies, as a research lead would record them after one call with each nominee.
| Question | Company L, weekly releases | Company M, quarterly releases |
|---|---|---|
| Release tracking since | 2021; changed tracker in 2023 | 2009, one tracker throughout |
| Releases shipped | About 190 since January 2023 | About 70 since 2009 |
| What each release leaves | Tickets and notes; tests automated, sign-off not recorded | Plan, tickets, test sign-off, notes |
| Outcomes linked | Incidents kept in a separate tool, not linked | Defects linked to the release that caused them, since 2010 |
| Who created the records | Employees, plus 30 contractors | Employees |
| Customer-owned material | None reported | Customer-specific builds belong to those customers |
| Full-time headcount at peak | 120 in 2026 | 210 in 2019; 180 today |
| Dated event | Old tracker read-only until April 2027 | None |
| CEO wants the conversation | Not asked | Yes, after the next release |
A launch screen picks L: roughly 52 releases a year against M's 4. The interview reverses the order for now. M holds seventeen years of connected release records, outcomes linked to their causes, and a CEO who wants to talk; its customer-specific builds are simply left out of scope. L has a dated event, the April 2027 switch-off of its old tracker, but a shorter history, unrecorded sign-offs and contractor work that SourceX's rights review would have to consider. Neither cadence lowers the bar both companies must clear: a peak of 50+ full-time employees, not counting contractors, a documented operating history, rights to license, and a go-ahead from the CEO. 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.
Does fast shipping leave operating history worth a conversation?
Five tests separate a cadence that matters from one that only looks impressive.
- Is the history older than the cadence? L's weekly rhythm began in 2023; question 1 asks what existed before it.
- Do the records connect? A plan, its issues, its tests, its sign-off and its outcome in one chain beats five unconnected logs. Whether one system can carry that chain alone is weighed in the one-system question.
- Are outcomes recorded? A rollback linked to its release is evidence of how work ended; an unlinked incident queue is not.
- Whose work is it? Contractors are excluded from the headcount test, customer-owned builds are usually not a fit, and outside contributions to a public repository were not created by the company's staff.
- Did a tool change split the history? The old tool's switch-off date is real timing, unlike a launch date.
Slow shippers get the same five tests. A company that ships four times a year can hold more of what matters than one that ships daily, and SourceX's process page says company size alone does not determine value: value depends on actual Acquisition Partner demand and due diligence. Partners who meet these records inside engineering engagements will find the consultant's angle in spotting referral opportunities during Jira projects, and the record types themselves under software engineering workflows.
Rank conversations by history, ownership and launch-week capacity
Put the finished interviews in order using six observable keys: a confirmed CEO path, a dated start to the history, connected records, clear ownership, a dated event, and capacity. Capacity means two things here: never raise this in a launch week, and keep your own research hours under a cap you set, since no benchmark has been published. Cadence comes last.
| Rank | Company (Illustrative) | CEO path | History starts | Records connected | Ownership | Dated event | Cadence |
|---|---|---|---|---|---|---|---|
| 1 | M, industrial controls software | Yes | 2009 | Yes | Own, minus customer builds | None | Quarterly |
| 2 | L, workflow software | Not asked | 2021 | Partly | Own; contractor work to review | Old tracker, April 2027 | Weekly |
| 3 | 70-person app studio building apps for client brands | Yes | 2017 | Yes | Client-owned | None | Daily |
| 4 | Portfolio company that launched its first product in 2025 | Founder keen | 2024 | Yes | Own | None | Weekly |
Row 3 drops out for now, because its fastest-moving records belong to its clients. Row 4 passes on ownership but has two years of history against a screening profile that looks for several years in operation; mark it revisit, with the year that would change it. L moves to first the day its CEO agrees, because April 2027 puts its older tracker history on a clock. The same columns sit comfortably beside a growth signal dashboard if your team keeps one.
What a launch calendar hides
- Cadence can be cosmetic. Feature flags and very small releases inflate counts without adding history.
- Public pages show what a company chooses to publish. Private products and repositories are invisible from outside.
- Deployment frequency measures automation as much as work. DORA's metric counts deployments; it says nothing about the records around them.
- Customer environments. Releases built into customers' systems, or owned by them, usually fall outside what a company can license.
- Board knowledge stays confidential. A product update in a board pack can start a question to the CEO, never a message to anyone else.
- Demand and programs move. SourceX checks relevance to current needs in review, a release history that misses one program's criteria may still suit a different program later, and broad analysis of AI data demand belongs on sourcex.si, not in this interview.
Method note. The interview applies SourceX's screening page, program terms (Version 2026-10-22.2, since replaced by version 2026-11-04), process page and release-coordination page, with DORA's metric definitions and GitHub's documentation of releases, all checked on 26 October 2026. Companies L and M, their releases and headcounts are invented, and the interview has not been tested on a real cohort; SourceX has published no conversion data. It covers what a company can say about its product history before any introduction; record quality, rights and demand wait for SourceX's review. To test it, interview the five fastest-shipping and the five slowest-shipping companies in your network and compare how many answer questions 4 to 6 fully and how many CEOs agree to a conversation. A company that wants to proceed applies on its own, through your partner link. Credit goes to the first valid introduction within the 90-day attribution window, once that company's verified application arrives, and the partner submitting it must already have signed the SourceX Channel Partner Agreement. The portfolio side is covered for venture platform teams, and related guides sit in the guides library.
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 interview shows a dated, connected product history the company owns and a CEO who wants the conversation, register for a partner link, compare the answers with who qualifies and read the program terms before anyone contacts the company.
A company's release history is a referral question; SourceX's own prospects as a business are an investment question for separate study, and anyone pursuing it can ask for SourceX's materials on its investor page.
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.