Record-to-Report: Period Close Process Mining Analysis

1. 0Executive summary
2. 1Method, lineage and assumptions
3. 2What the account records — and does not
4. 3Period close: state and cycle time
5. 4Journal population and process model
6. 5Reversal, reclassification and approval trail
7. 6Who touched what
8. 7Intercompany and elimination
9. 8Exception registers
10. 9Data quality
11. 10Findings and recommended actions
12. AAppendix A — SuiteQL used
13. BAppendix B — Reducer logic
14. CAppendix C — Worked hand-check
15. DAppendix D — Glossaries and run metadata

0Executive summary

Record-to-report is the process finance leads most want measured and least often see. In this account it is not so much unmeasured as absent: no accounting period ending after November 2021 has been closed or locked. The 71 periods that were closed (Jan 2016 – Nov 2021) were closed in four bulk batches on four calendar days, a median of 748 days after period end — the shortest interval was 71 days, the longest 1,417. Every one of the 25 periods carrying the account's live transaction history (Oct 2024 → Sep 2026, 6,240 posting transactions, $19.8M gross in 2026 alone) is open to posting with A/R, A/P and GL unlocked, and the oldest unclosed period, Dec 2021, has been open 1,708 days.

Three findings follow from what can be measured:

1. The close is a state, not a process. NetSuite records no system note when a period is locked or closed in this account; only closedondate survives. Cycle time from period end to close is therefore known for 71 historical periods (median 748 days, in four batches) and undefined for the 57 open ones. There is no close checklist activity, no lock sequence (A/R → A/P → All), and nothing that distinguishes a soft close from a hard one. §3.
2. Journals are few, but a third of them are reversals or tests. Of 40 genuine journal entries, 6 pairs (12 entries, 30%) are original + reversal, 5 (12.5%) are reclassifications or allocations, and 9 (22.5%) carry a "Test" memo or an Atlas sweep-test external id and post $100–$555 to real bank and A/R accounts in the current period. Four entries ($10,316) sit in Pending Approval with no approver assigned, two of them in periods that have already ended. Only one approval transition is observable in system notes, and it was self-approved by the administrator who entered it. §5.
3. Intercompany runs, elimination has never. Three Advanced Intercompany Journals (JE91–JE93, $1,000/month, Sub 2 → Sub 3) and six intercompany operational documents exist; the elimination subsidiary xElim has zero lines in its history, the two accounts flagged for elimination (1910, 2910) have zero postings, and no transaction line anywhere carries eliminate = T. Consolidated statements have never been eliminated. §7.

1Method, lineage and assumptions

The O2C and P2P studies used documents as objects and link tables as edges. Record-to-report has a different shape: the case is the accounting period; the events are (a) postings into the period, (b) journal entries with their approval, reversal and reclassification lineage, and (c) lock/close actions on the period record. Events of type (c) are read from systemnote; they turned out not to exist (§2), so the period lifecycle collapses to its recorded end state.

Assumptions register

2What the account records — and does not

3Period close: state and cycle time

Figure 1 — All 168 monthly periods. Emerald = closed (with days from period end to close); grey = open past end (bar = days open as at 2026-09-04); hollow = future. Four close batches are visible as plateaus.

The pattern is unambiguous: periods were closed in bulk, years in arrears, on four occasions — and not at all since 2022-02-09. All 71 closed periods have alllocked = aplocked = arlocked = T; all 97 open periods have every lock flag F and allownonglchanges = F. There is no soft-close (locked but not closed) state anywhere. The 25 periods that hold the account's operational history are all posting-open:

4Journal population and process model

Figure 2 — The 40 genuine journals classified by purpose (A4–A6). Counts and debit totals.

Journal lifecycle model

Figure 3 — Journal lifecycle as observed. Solid = evidenced by a system field or system note; dashed = inferred from header state only. Counts are journals.

The reference lifecycle is Created → (Pending Approval → Approved) → Posted → [Reversed]. What the account evidences: 40 created (9 with a create note, 31 without); 4 currently Pending Approval; 1 observed approval transition (JE100); 36 posted; 6 reversed by a linked entry. There is no rejection, no edit-after-post, no reopen, and no voided journal. Because the approval step leaves a system note only when it happens through the UI and every other journal arrived Approved for Posting at creation, the process is effectively Created-as-Approved → Posted for 35 of 40 entries.

5Reversal, reclassification and approval trail

Reversal pairs (6)

Reversal rate: 6 of 40 genuine journals are reversals (15.0%); 12 of 40 (30.0%) participate in a reversal pair. Excluding test entries, 3 of 31 substantive journals are reversals (9.7%), and all three are legitimate accrual or void reversals. Median original → reversal interval: 1.5 days overall; 20 days for the one genuine month-end accrual (JE49 → JE50). Half of the reversal pairs are invisible to a search on the reversal field — the Atlas-generated reversals set memo and externalid but not the system reversal link.

Reclassification and allocation (5)

Approval trail

6Who touched what

Figure 4 — Genuine journals by last-modifying identity (bars) and by creator where recorded (dots).

Segregation of duties cannot be demonstrated from the record: no journal shows a creator, approver and poster as three different people; the single observed approval was performed by the entrant; the finance roles (Controller, Accountant) touch one journal each while a system-aliased identity carries 67.5% of entries, including every entry over $10,000.

7Intercompany and elimination

8Exception registers

Complete lists. Ages as at 2026-09-04.

8-A · Open periods past end (57) — summary by year

8-B · All genuine journals (40)

8-C · Test entries posted to production (9)

JE163 (Dr 1100 A/R / Cr 5340 COGS, $100, Parent Company), JE164 (Dr 1100 A/R / Cr 1090 Undeposited Funds, $555, Parent Company), JE165 (petty cash ↔ checking, $100), JE166–JE171 (1014 Payroll ↔ 1010 Checking, $100 each, three pairs net to zero). Net residual on the ledger: JE163 + JE164 + JE165 = $755 of unreversed test postings — $655 of it inflating Parent Company A/R and understating COGS / Undeposited Funds in Sep 2026. All nine were entered by the administrator between 2026-05-08 and 2026-05-17 (system-note clock) into Sep 2026 trandates.

8-D · Intercompany documents (9) — see §7 table

9Data quality — what limits the analysis

10Findings and recommended actions

1. Institute the close — start by locking, not closing. Fifty-seven periods past their end are open to posting; the 25 that carry live history (Oct 2024 → Aug 2026, $19.8M gross in 2026) can be altered by anyone with transaction permissions and no note would record it. Lock A/R, A/P and All on every period through Jul 2026 now; close Dec 2021 → Dec 2024 (zero or complete activity) outright; then adopt a standing rule — lock by working day 5, close by working day 10 — and enforce it through the Period Close Checklist so the close leaves a trail. Measure cycle time from the first month this is done; today it cannot be measured because it does not occur.
2. Clean the journal population and close the approval gap. Reverse JE163, JE164 and JE165 ($755 of test postings sitting in Parent Company A/R, COGS, Undeposited Funds and petty cash — the six Atlas pairs already net to zero). Assign an approver and clear the four pending journals ($10,316; JE99 and JE101 are in a period that ended 35 days ago). Require the reversal link on script-generated reversals so all six pairs — not three — are discoverable. Enforce a memo on every journal: JE88, JE89, JE91–93, JE100–101, JE159–162 ($292K) have none.
3. Run elimination once, then every month. The elimination subsidiary and eliminate-flagged accounts exist and have never been used; JE91–93 are the only intercompany postings and they self-net, so a first elimination run will post nothing material — which is exactly the right moment to establish the step. Pair it with the actor finding: two-thirds of journals, including all eight capital injections ($680K) and three inventory reclasses ($262K, no memo, created the same day), are attributed to a negative-id system identity that displays as a person; the Controller has touched one journal. Decide who owns journals in this account and make the record show it.

AAppendix A — SuiteQL used

Verbatim as executed. Q1–Q2 feed the reducer; the rest are probes and registers.

Q1 — Monthly periods with close state (reducer input periods)

SELECT id, periodname, TO_CHAR(startdate,'YYYY-MM-DD') AS startdate, TO_CHAR(enddate,'YYYY-MM-DD') AS enddate, closed,
       TO_CHAR(closedondate,'YYYY-MM-DD') AS closedon, alllocked, aplocked, arlocked
FROM accountingperiod WHERE isquarter='F' AND isyear='F'

Q2 — Period-bearing transactions (reducer input txns)

SELECT t.id, t.type, t.tranid, TO_CHAR(t.trandate,'YYYY-MM-DD') AS trandate, TO_CHAR(t.createddate,'YYYY-MM-DD') AS created,
       t.postingperiod, t.posting, t.status, t.journaltype, t.isreversal, t.reversal, t.createdby, t.lastmodifiedby,
       CASE WHEN t.memo LIKE 'Beg Balance%' THEN 'T' ELSE 'F' END AS begbal,
       CASE WHEN t.memo LIKE 'Reversal of%' OR t.externalid LIKE '%reversal-of%' THEN 'T' ELSE 'F' END AS memo_reversal,
       CASE WHEN t.externalid LIKE 'Reclass%' OR t.memo LIKE '%Reclass%' OR t.memo LIKE '%Allocation%' THEN 'T' ELSE 'F' END AS reclass
FROM transaction t
WHERE t.postingperiod IS NOT NULL
  AND t.type NOT IN ('SalesOrd','PurchOrd','Estimate','Opprtnty','WorkOrd','TrnfrOrd','VendAuth','RtnAuth','PurchReq')

Q3 — Journal census

SELECT t.type, COUNT(*) AS docs, COUNT(DISTINCT t.status) AS statuses, TO_CHAR(MIN(t.trandate),'YYYY-MM-DD') AS first_date,
       TO_CHAR(MAX(t.trandate),'YYYY-MM-DD') AS last_date,
       SUM(CASE WHEN t.memo LIKE 'Beg Balance%' THEN 1 ELSE 0 END) AS beg_bal, SUM(CASE WHEN t.posting='T' THEN 1 ELSE 0 END) AS posting
FROM transaction t WHERE UPPER(t.type) LIKE '%JOURNAL%' GROUP BY t.type

Q4 — System notes on journals (field coverage)

SELECT sn.field, sn.type, COUNT(*) AS notes, COUNT(DISTINCT sn.recordid) AS journals, COUNT(DISTINCT sn.name) AS actors
FROM systemnote sn JOIN transaction t ON t.id = sn.recordid AND sn.recordtypeid = -30
WHERE t.type = 'Journal' GROUP BY sn.field, sn.type ORDER BY COUNT(*) DESC

Q5 — System notes on accounting periods (none found)

SELECT sn.recordtypeid, sn.field, COUNT(*) AS notes, COUNT(DISTINCT sn.recordid) AS recs
FROM systemnote sn WHERE UPPER(sn.field) LIKE '%LOCK%' OR UPPER(sn.field) LIKE '%CLOSE%'
   OR UPPER(sn.field) LIKE '%ACCTPERIOD%' OR UPPER(sn.field) LIKE '%ACCOUNTINGPERIOD%'
GROUP BY sn.recordtypeid, sn.field ORDER BY COUNT(*) DESC

Q6 — Genuine journals with lineage fields

SELECT t.id, t.tranid, TO_CHAR(t.trandate,'YYYY-MM-DD') AS trandate, TO_CHAR(t.createddate,'YYYY-MM-DD') AS created,
       ap.periodname, BUILTIN.DF(t.status) AS status_label, t.posting, t.isreversal, t.reversal,
       TO_CHAR(t.reversaldate,'YYYY-MM-DD') AS reversaldate, t.journaltype, t.createdby, t.lastmodifiedby,
       BUILTIN.DF(t.createdby) AS created_by, t.externalid, t.memo
FROM transaction t LEFT JOIN accountingperiod ap ON ap.id = t.postingperiod
WHERE t.type = 'Journal' AND (t.memo IS NULL OR t.memo NOT LIKE 'Beg Balance%') ORDER BY t.trandate, t.id

Q7 — Journal line composition

SELECT t.id, t.tranid, tl.subsidiary, a.accttype, COUNT(*) AS lines,
       ROUND(SUM(CASE WHEN tal.amount > 0 THEN tal.amount ELSE 0 END),2) AS debits,
       ROUND(SUM(CASE WHEN tal.amount < 0 THEN -tal.amount ELSE 0 END),2) AS credits,
       MAX(a.acctnumber || ' ' || a.fullname) AS sample_account
FROM transaction t JOIN transactionline tl ON tl.transaction = t.id
JOIN transactionaccountingline tal ON tal.transaction = t.id AND tal.transactionline = tl.id JOIN account a ON a.id = tal.account
WHERE t.type = 'Journal' AND (t.memo IS NULL OR t.memo NOT LIKE 'Beg Balance%')
GROUP BY t.id, t.tranid, tl.subsidiary, a.accttype ORDER BY t.id, tl.subsidiary, a.accttype

Q8 — Full system-note trail on journals

SELECT t.tranid, sn.recordid, TO_CHAR(sn.date,'YYYY-MM-DD HH24:MI') AS note_time, sn.name AS actor_id, BUILTIN.DF(sn.name) AS actor,
       sn.field, sn.type, sn.oldvalue, sn.newvalue, sn.context, BUILTIN.DF(sn.role) AS role
FROM systemnote sn JOIN transaction t ON t.id = sn.recordid AND sn.recordtypeid = -30
WHERE t.type = 'Journal' ORDER BY sn.date, sn.recordid

Q9 — Pending-approval queue

SELECT t.id, t.tranid, TO_CHAR(t.trandate,'YYYY-MM-DD') AS trandate, TO_CHAR(t.createddate,'YYYY-MM-DD') AS created, ap.periodname,
       t.journaltype, t.nextapprover, BUILTIN.DF(t.nextapprover) AS approver, t.approvalstatus, BUILTIN.DF(t.lastmodifiedby) AS last_modified_by,
       t.memo, (SELECT ROUND(SUM(CASE WHEN tal.amount > 0 THEN tal.amount ELSE 0 END),2) FROM transactionaccountingline tal WHERE tal.transaction = t.id) AS debit_total
FROM transaction t LEFT JOIN accountingperiod ap ON ap.id = t.postingperiod WHERE t.type = 'Journal' AND t.posting = 'F' ORDER BY t.trandate

Q10 — Intercompany / elimination evidence

SELECT (SELECT COUNT(*) FROM transactionline WHERE eliminate = 'T') AS elim_flagged_lines,
       (SELECT COUNT(*) FROM transactionline WHERE subsidiary = 4) AS sub4_lines,
       (SELECT COUNT(*) FROM account WHERE eliminate = 'T') AS elim_accounts,
       (SELECT COUNT(*) FROM transaction WHERE intercompany = 'T') AS interco_flagged_txns,
       (SELECT COUNT(*) FROM transaction WHERE intercotransaction IS NOT NULL) AS interco_paired_txns,
       (SELECT COUNT(*) FROM accountingperiod WHERE isquarter='F' AND isyear='F' AND closed='F' AND enddate < SYSDATE) AS open_periods_past_end,
       (SELECT MIN(periodname) FROM accountingperiod WHERE isquarter='F' AND isyear='F' AND closed='F') AS oldest_open_period
FROM dual;

SELECT a.id, a.acctnumber, a.fullname, a.accttype, a.eliminate, (SELECT COUNT(*) FROM transactionaccountingline tal WHERE tal.account = a.id) AS postings
FROM account a WHERE a.eliminate = 'T' OR a.fullname LIKE '%Intercompany%' OR a.fullname LIKE '%Elimination%' ORDER BY a.acctnumber;

SELECT t.id, t.type, t.tranid, TO_CHAR(t.trandate,'YYYY-MM-DD') AS trandate, BUILTIN.DF(t.status) AS status_label, t.intercompany, t.intercotransaction,
       BUILTIN.DF(t.entity) AS entity, t.tosubsidiary, ROUND(ABS(t.foreigntotal),2) AS total
FROM transaction t WHERE t.intercompany = 'T' OR t.intercotransaction IS NOT NULL ORDER BY t.trandate, t.id

Q11 — Trandate vs posting period mismatch (0 rows)

SELECT t.type, COUNT(*) AS docs FROM transaction t JOIN accountingperiod ap ON ap.id = t.postingperiod
WHERE t.posting = 'T' AND (t.trandate < ap.startdate OR t.trandate > ap.enddate) GROUP BY t.type

Q12 — Close batches

SELECT TO_CHAR(closedondate,'YYYY-MM-DD') AS closed_on, COUNT(*) AS periods, MIN(periodname) AS first_period, MAX(periodname) AS last_period,
       MIN(closedondate - enddate) AS min_days, ROUND(AVG(closedondate - enddate),0) AS avg_days, MAX(closedondate - enddate) AS max_days
FROM accountingperiod WHERE isquarter='F' AND isyear='F' AND closed='T' GROUP BY TO_CHAR(closedondate,'YYYY-MM-DD') ORDER BY 1

BAppendix B — Reducer logic

// sqlReduce, contract "whole": rows.periods (168), rows.txns (6,292)
P = indexBy(periods, id); today = 2026-09-04
for each period: rs = txns in period; posting = rs.posting==='T'; je = type==='Journal'; realJe = je && begbal==='F'
  reversalJe = realJe with isreversal==='T' or memo_reversal==='T'; reversedJe = realJe with reversal set and not itself a reversal
  reclassJe = realJe.reclass==='T'; icJe = journaltype contains 'INTERCOMPANY'; pendingJe = realJe.posting==='F'
  daysOpen = closed ? null : daysBetween(enddate, today); closeDays = closedon ? daysBetween(enddate, closedon) : null
closeCycleDays = H.stats(closed periods' closeDays); closeBatches = countBy(closedon)
openPastEnd = open periods with enddate < today → H.stats(daysOpen)
actor matrix = countBy(realJE, lastmodifiedby) and countBy(realJE, createdby); reversalPairs from isreversal rows; memoReversals from memo flag
return per-period rows for Oct 2024 → Oct 2026 + aggregates (6.6 KB)

Worker 18 ms; queries 944 ms; 1.75 MB decoded / 147 KB on wire. Classification of the 40 journals into purpose classes (§4) was done from Q6/Q7 output by inspection — the population is small enough that every entry is listed in §8-B.

CAppendix C — Worked hand-check

DAppendix D — Glossaries and run metadata