How to build a case chronology from emails and documents

Your supervisor needs to know what happened, when, and where the account comes from. A useful working chronology lets them check each entry without reconstructing your research. It also shows what the documents do not establish.

This guide is for a junior disputes paralegal preparing an internal chronology for a supervising lawyer. It teaches evidence organisation, not liability, admissibility, privilege, filing or legal deadlines. You need a spreadsheet or text-table editor and permission to work with the source documents. No specialist paid product is required for the example.

The quick task path

1. Agree the question and authorised source set with the supervising lawyer.
2. Give every source a stable ID; identify copies without deleting the originals.
3. Record events using their event dates, with exact source locators.
4. Attribute accounts, retain contradictions and label plans as plans.
5. Keep ranges and unknown dates visible; explain any time conversion.
6. Check every row against its sources and hand over the open questions.

See the result first

Juniper Events ordered display panels from Beacon Print. The fictional matter takes place in June 2026. The supplied records disagree about delivery time and quantity. A few rows from the completed chronology show the approach. Times in this table use the stated local offset, UTC+03:00; a stated time is not a finding that the event happened exactly then.

The complete example contains thirteen rows. It is one defensible structure, not the only acceptable wording or way to group an inspection’s observations.

1. Set the question and preserve the starting point

For this exercise, the question is: What do the supplied records say about the delivery, quantity, condition and follow-up, and what remains unresolved? Record the scope and cut-off in the instruction sheet. Include material that weakens as well as supports Juniper’s account.

Work from the supplied files, retain them unchanged and make working copies where needed. In a real matter, ask the supervising lawyer about preservation, confidentiality, access and any restrictions before copying, uploading, deleting or sharing material. Use the fictional packet while those questions are unresolved. This worksheet does not establish a chain of custody or authorise disclosure.

2. Index the documents before extracting events

The source index records the ID, file location, stated document date, creator and limitations. These are dates within the records, not the time your computer downloaded the files.

D10 is a byte-identical copy of D05, received through another route in the fictional packet. Keep the second intake entry and mark its relationship to D05. Do not turn two copies into two independent accounts of delivery. The thirteen-row example therefore does not cite D10 as corroboration.

Keep IDs stable when new material arrives. Do not renumber documents or events just because their display order changes. Every paragraph in the synthetic files has a locator such as D06-P02; use that exact locator instead of “see emails.”

3. Extract the event, not just the document’s date

Read each relevant passage and write a short, neutral description. Capture enough detail to distinguish the occurrence, its source and its limits. One source can support several events; one event may need several sources.

D03 is an email dated 12 June at 09:45, reporting a van departure at 09:30. The row uses 09:30 as the reported event time and retains 09:45 in source_dates.

D07 was forwarded on 13 June at 07:20 UTC, equivalent to 10:20 at UTC+03:00. Its quoted message is dated 12 June at 15:10 UTC+03:00. E12 belongs at the quoted message time, with its indirect source made explicit. Neither timestamp proves that the proposed panels arrived.

The workbook separates event_start, event_end, date_precision and source_dates. Entry cells use Text format, and the completed example includes a UTC sort helper for timed entries. The helper is a fixed value, not a formula: update or clear it when an event time changes. It orders the display; it does not verify a document’s clock or establish which disputed account is true. Check pasted or imported values and see the field guide for safe sorting and missing-offset cases.

4. Keep attribution, disagreement and uncertainty visible

Prefer “Mara reports counting 16 panels at 14:00” to “Only 16 panels were delivered.” D06 says she did not count unloading and could not exclude panels being moved elsewhere. The count and the delivery quantity are different questions.

Similarly, keep all of these visible: D05’s 11:40 receipt entry, D04’s gate record, Mara’s noon observation and Imani’s approximate recollection. Do not select the version that best supports the client’s position or silently combine them into one certain delivery time.

Use basis to describe what supports the entry: a document record, an attributed report, a disputed record, a scheduled milestone, an approximate recollection or a quoted intention. These are working labels, not legal findings or statistical confidence scores.

For D08, keep the 12:15–12:25 interval. Do not invent a 12:20 midpoint. Keep the later call in an undated group; do not assign it the note’s file timestamp. Sort ranges by their earliest stated bound only for display and retain the full interval. Where ranges overlap other entries, their exact relative order remains unresolved.

5. Check, hand over and update

Read each row back against its cited passage. Check dates, offsets, attribution, quantities and missing context. Reconcile the source index, duplicate labels and open questions. Ask a second person to locate several events from your references without help; that checks usability, not legal correctness.

A short handover could say:

The receipt records 18 panels at 11:40, while the gate and buyer accounts differ. Mara’s 14:00 count does not establish the unloading quantity. We have only a forwarded intention concerning further panels, and one call remains undated. The next questions concern the receipt’s maker, the gate clock, the complete inventory and evidence of any later supply.

Keep the owner of each question explicit. “Pending legal interpretation” in a cell is a reminder to obtain instructions; it is not a technical lock that prevents action.

When new evidence arrives, add its source ID and review the affected rows. Use the workbook’s Change log sheet or the blank CSV change log. Record each changed field’s old value, new value, reason and source locator, including changes to references, qualifications and the UTC helper. For a new event, record each nonblank field. Preserve the earlier source and chronology snapshot; the log complements those records rather than replacing them. A corrected gate record does not automatically settle the delivery receipt, quantity or condition. The separate changed-input task provides a transfer exercise; in an observed test, the facilitator supplies it only after the baseline attempt.

Common mistakes

Before handing over, use the final checklist. The goal is a checkable working account, including the questions that remain open.