Specialist advisors and industry operators are usually the first to admire a postmortem habit. A delivery consultant who ran a client's retrospectives, an engineering leader who wrote the template, a construction executive whose closeout reviews are the envy of peers: each has seen the archive work. That admiration is the reason to slow down, because the record type never establishes rights, fit with a program or any buyer's interest.
A postmortem, described by its parts
A postmortem is the written review a team produces after a project, release, incident or phase ends. Companies call it a retrospective, a lessons-learned review, an after-action review, a closeout report or a post-incident review. The 2020 Scrum Guide describes the Sprint Retrospective as the point where a team plans how to improve quality and effectiveness, looking back at its people, interactions, processes, tools and definition of done, what went well and which problems came up. Project-level reviews ask the same questions over months instead of weeks.
Most write-ups share a skeleton: a summary, a timeline, the causes the team identified, what worked, what did not, and follow-up actions with owners. The value sits in the links between those parts, a decision, its outcome and the change it triggered, repeated across dozens of projects. Project histories and project histories with decisions and outcomes cover what makes such records useful; the main site's look at a construction project archive covers what a closed project leaves behind. This page covers the step before either: whose words are on the shelf.
Why the celebration comes too early
Three habits of good postmortem cultures cut against licensing.
- They are written about other people's projects. Agencies, consultancies, contractors and implementation teams run most of their projects for clients. A candid review of a client project describes the client.
- They borrow freely. To explain a cause, teams paste in whatever proves it: a vendor's incident report, a subcontractor's delay notice, a client's email, a set of drawings.
- They record people. Blameless templates aim at systems rather than individuals, yet many reviews still name who was on call, who approved what and who missed a handoff.
SourceX's screening guidance names project histories among the original documentation it looks for, and in the same breath treats material that belongs to clients as usually not a fit, requiring rights that do not breach client confidentiality, employee privacy or third-party contracts. A postmortem archive can satisfy the first sentence and fail the second.
The postmortem ownership question sheet
Nine questions in three groups. An operator who knows the archive can answer every one without opening a document.
| # | Question | Answer that points to the company | Answer that raises a rights question | Who can answer |
|---|---|---|---|---|
| 1 | Who wrote the write-ups? | Employees | Contractors, consultants or a client's staff | The template owner or PMO lead |
| 2 | Did anyone outside co-write or sign off? | No | A client approved them; a joint-venture partner co-authored | The project lead |
| 3 | Where are they kept, under whose account? | The company's own wiki or drive | A client's portal or a shared project platform | The system administrator |
| 4 | Whose project was it? | An internal project, product or operation | Delivered for a client under contract | The delivery or operations lead |
| 5 | What does the client contract say about project documentation? | Not a client project, or the company keeps its working papers | The client owns deliverables, or nobody knows | The company's contracts owner, never the introducer |
| 6 | How much does a write-up describe the client's business? | Little beyond a project code | Strategy, systems, results or incidents at the client | The project lead |
| 7 | What is pasted in from others? | Nothing, or only the company's own logs and notes | Drawings, vendor incident reports, subcontractor notices, client emails | The template owner |
| 8 | Are people named? | Roles only | Individuals, with performance comments | HR or the privacy lead, in categories |
| 9 | Has the archive been licensed, shared or published before? | Never | Licensed for AI training, or shared under a client's confidentiality terms | The CEO or finance lead |
Questions 1 to 3 are about authorship, 4 to 6 about the project, 7 to 9 about contents. Count the answers in the rights-question column for each group. The CEO hears the sorted picture, for example "about seven in ten were written by staff about internal work with nothing pasted in", not "we have a great archive".
Question 9 has a hard edge. Records already licensed for AI training are out, because in SourceX's process the supplier confirms the data has not been licensed for that purpose before. And if the company itself has ever applied, held an account or exchanged email with SourceX, section 3 of the terms makes it a duplicate that no later introduction is credited for.
Construction shows the pasted-in problem plainly. Under the AIA's B101-2017 owner-architect agreement, as AIA Contract Documents summarizes it, the architect and its consultants are treated as the authors and owners of their Instruments of Service and keep their rights, including copyright, while the owner receives a nonexclusive license to use them for the project. A contractor's closeout review that reproduces sheets of the drawings therefore carries another firm's work inside the contractor's own analysis, and the RFI log page shows how design questions pass between firms on one project.
Method note. The sheet applies SourceX's screening profile and program terms (version 2026-10-22.2, since replaced by version 2026-11-04), SourceX's process description of what a supplier confirms, the 2020 Scrum Guide's description of a retrospective and AIA Contract Documents' summary of B101-2017, all checked 2026-10-27. The three groups and nine questions are ours. The sheet does not assess demand, record quality or value, and it cannot interpret a contract; SourceX's rights review and the company's counsel do that.
Original work, mixed content and an open question (Illustrative)
Illustrative. Generic companies, invented counts estimated by each owner, no outcome implied.
A. A 380-person business software company: mostly original. Its engineering and product teams have written 210 postmortems since 2017, on incidents and launches, in the company's own wiki. The VP of engineering estimates that 196 were written by employees and 14 by an outside reliability consultancy in 2018. Of the 196, about 46 paste in a cloud provider's incident report or customer emails. That leaves 150 written by staff about the company's own systems with nothing borrowed, the set the CEO can hear about first. The other 60 wait for rights review, and customer account IDs in impact sections are a privacy item either way.
B. A 190-person general contractor: mixed third-party content. It keeps 64 closeout reviews from 2012 to 2025. The lessons-learned narrative in each is the contractor's own, but every review attaches drawing sheets, subcontractor delay notices and the owner's meeting minutes. The route is to describe the contractor's own narratives and say plainly that the attachments belong to others and are likely outside any discussion.
C. A 95-person digital agency: ownership unclear. It has 140 campaign retrospectives. About half were written by freelance strategists, all describe client campaigns, and most include the client's own results. Whether the agency may license any of it depends on client contracts and freelancer agreements nobody has read with this question in mind. The agency's 20 internal process retrospectives are a separate, cleaner set. The route is to hold the campaign archive and mention only the internal set, if the CEO wants a conversation at all. Separating an agency's own records from client work goes further on that split.
Who ran the project, who wrote the review, who decides now
- Name the entity that ran the projects. If the company has since been acquired, the archive sits wherever the deal put it, and the acquirer's CEO now decides for it.
- Sort authors by type: employees, contractors, client staff, partners. A single contractor-written year can carry a different answer from the rest.
- For client projects, find who holds the contract and whether it addresses project documentation. The company reads it; you never interpret it.
- Trace pasted material to its source: vendor reports, drawings, client correspondence, subcontractor notices.
- Ask who owns the archive today inside the company: the PMO, engineering, quality. That person can describe it and sponsor the idea internally. The CEO authorizes a licensing discussion.
The fit checks run in parallel: 50 or more full-time employees at the company's peak, contractors excluded; several years of records it created itself; rights it can grant; and a CEO who wants the conversation. Size and geography follow who qualifies. A company also needs history beyond the postmortems themselves, and SourceX's qualification asks about systems and years, never for files.
Seven questions for the first call, and four requests never to make
Each can be answered from memory:
- Roughly how many postmortems exist, and from which year?
- What kinds of project or incident do they cover?
- Where do they live, and who administers that system?
- What share were written by employees?
- What share concern client projects?
- Do they usually attach outside material, and what kind?
- Would the CEO want to hear about licensing the company's own write-ups?
Never ask for a sample postmortem, the filled-in template, a read-only link to the wiki or a list of client projects. An advisor who facilitated retrospectives keeps what they saw inside that engagement. A first note to the CEO might read: "Your teams have written about 200 project reviews since 2017, most by staff about your own systems. SourceX licenses company-owned history like this for AI training and evaluation, after its own rights review. Would a conversation be useful? Nothing from the archive has been shared." The product change log guide sorts another kind of authored history the same way.
Limits of the question sheet
- Answers come from memory. Owner estimates can be off by a wide margin, and a clean author answer can hide a pasted vendor report.
- Blameless is not anonymous. Templates that avoid blame still carry client detail and customer identifiers.
- Contracts decide, not the sheet. Whether a client agreement leaves the company free to reuse its working papers is a legal reading.
- Demand is unknown. SourceX checks relevance to current data needs in review, and a well-sorted archive can still be declined.
- Published reviews are a different set. Write-ups already released as case studies are public, add little to a licensing discussion and say nothing about rights in the unpublished rest.
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 the sheet shows a set of staff-written reviews about the company's own work and the CEO wants to talk, register for a partner link, confirm the basics on who qualifies and read the program terms before the note is sent.