This packet is part of the data-break-to-repair records guide. It describes the same chain that the before-and-after case shows in action, and it pairs with the seven-check grading sheet if you want to grade the chain as well.
The packet, field by field
Ten fields. Each is answered in words, a year or a band, and each names the person who confirmed it.
| Field | What to write | Answer format | Usually confirmed by |
|---|---|---|---|
| 1. Warehouse | The warehouse product and the year it went live | Name and year | Data platform owner |
| 2. Transformation framework | The framework that builds and tests models, managed or self-hosted, and since when | Name and year | Analytics engineering lead |
| 3. Tracker | Where data incidents are ticketed, and the project's start year | Name and year | Tracker administrator |
| 4. Code history | Whether the transformation project has been in version control for the whole period | Yes, partly or no, with year | Analytics engineering lead |
| 5. Object types | Which kinds of records exist: test outcomes, run logs, tickets, code changes, warehouse query logs | List | Analytics engineering lead |
| 6. Volume band | Rough scale: data tickets per year, runs per day | Band, never an exact count | Tracker administrator |
| 7. Join keys | What ties one record to the next: a test identifier, a ticket reference in code changes | Key and since when | Analytics engineering lead |
| 8. Run artifact retention | Whether per-run results are archived or only the latest survives | Archived since a year, or not | Analytics engineering lead |
| 9. Verification record | Where finance or the report owner signs off a corrected number | System and year | Controller or finance lead |
| 10. Unknowns | Anything nobody present could answer | Short list with who might know | Whoever is closest |
Fields 7 and 8 do most of the work. The transformation framework's per-run output file records each test that ran with its status and a count of failing rows, so it is where a past failure is written down. If those files were not kept, the failure has no record of its own. In the tracker, a code change is tied to a ticket when its message quotes the ticket key, so the share of changes that quote one tells SourceX how much of the chain joins without anyone reading text.
Field 1 has a time limit worth noting. The warehouse's own execution log is not permanent: Snowflake's account usage history of statements covers the last 365 days, so older execution history exists only if the company copied it somewhere.
An Illustrative filled packet
Illustrative. A fictional 450-person specialty insurer with a six-person analytics engineering team.
| Field | Entry |
|---|---|
| 1. Warehouse | Snowflake, live since 2019 |
| 2. Transformation framework | dbt Core since 2020; stored procedures before that |
| 3. Tracker | Jira Software; data project since 2020 |
| 4. Code history | Model code in Git for its whole life |
| 5. Object types | Test outcomes, run logs, tickets, code changes, query logs |
| 6. Volume band | A few hundred data tickets a year; nightly and hourly runs |
| 7. Join keys | Test identifiers in run output; ticket key in most code changes since 2022 |
| 8. Run artifact retention | Archived per run since 2024; latest run only before |
| 9. Verification record | Finance sign-off as ticket comments since 2023; email before |
| 10. Unknowns | Whether warehouse tables include policyholder data under contract limits (general counsel); whether query history before last year was ever copied (data platform owner) |
Read as a whole, the packet says: from 2024 the company can likely join a failed test to its ticket, fix, rerun and sign-off; from 2022 to 2024 the ticket-to-fix link holds but the original failure may be missing; before 2022 the chain depends on free text. That is a useful, honest description. It does not promise a buyer, a value or a timeline.
What never goes in the packet
- Query text, SQL from models, or screenshots of either.
- Table, schema or model names that reveal clients, products or policyholders.
- Ticket titles, descriptions or comments, even paraphrased.
- Row counts from specific tables, test failure details, or sample rows.
- Names of individuals other than the role that confirmed each field.
- Anything the sponsor offers to "just send over." Decline it and point them to SourceX.
The same rule runs through the business data inventory worksheet and the metadata-only Jira inventory template. For a broader inventory across systems, see how to prepare a metadata-only inventory of business systems.
How to fill it in without seeing anything
Start with the analytics engineering lead and work down the table. Ask for years and bands, not lists. "Roughly how many data tickets a year?" gets a band; "Can you pull the count?" invites an export you do not want. If the answer is "I would have to check," write unknown and the name, and move on.
Ask the controller about field 9 separately. Data teams often believe finance signs off in Jira when it actually happens by email, or the reverse. A one-line answer from the person who signs is worth more than an assumption.
If lineage comes up, keep it at the level of "which tool records which link." The difference between tracing how data moved and proving where it came from is covered in data lineage vs data provenance, and it is a question for review, not for the packet.
Handing the packet to SourceX
The people who fill in the packet do not authorize a licensing discussion. That is the CEO, usually advised by the CFO, COO or general counsel, because warehouse tables can include customer records under contract terms the data team does not control. Share the packet only with the CEO, those advisers and SourceX.
Before you submit, sign the Channel Partner Agreement and confirm the baseline: 50+ full-time employees at peak (contractors excluded), documented operating history, rights to license, and an authorized contact. 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. Attribution goes to the first valid introduction within the 90-day window, credited when the company submits a verified application, with one partner per company.
What stays unknown until SourceX reviews it under the company's own agreement: whether the rights cover the records, how much of each era survives, and whether a buyer wants it. SourceX does not claim a connector to Snowflake, dbt, Jira or any other tool named here. More record locations are listed under systems.
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.
If you can fill in all ten fields, unknowns included, and the CEO will take the call, register as a partner, check the company against who qualifies, and read the terms before you submit the packet.