Data quality
What does not fit, you see first.
planist checks every data state against 31 rules in six areas. Every finding comes with its cause and the fix, and you decide whether it is resolved or explained. The check happens before the report, not in the meeting.
Data quality
last calculated 3 July, 14:02
- no findings
Budget value without any actual booking
Budget and actuals
Actual bookings without any budget value
3 accounts · 4,000 in total
- no findings
Months with quarterly granularity only
Import and coverage
- explained
Inactive account with figures
1 account · marked as explained
planist never corrects silently.
What you entered, planist does not change on its own account. Where it does decide something itself, such as an estimated account type, a split quarterly value or an AI suggestion, that shows up as a finding you can overrule. What is wrong stops and says why.
Six areas, 31 rules.
- 7 rules
Budget vs. actuals
Actuals without budget and budget without actuals, quarterly-only values, unconfirmed AI estimates in the active budget.
- 10 rules
Chart of accounts and master data
Origin of account properties, dimension coverage, account number format, duplicate account numbers, inactive accounts with figures, open planned accounts.
- 6 rules
Balance sheet and balances
Missing or late opening balances, balance sheet identity, opening balance on an income statement account, liquid funds without an anchor.
- 3 rules
Accounting purpose
Changes between years, unconfirmed or empty purposes, accounts that diverge between the views.
- 4 rules
Import and coverage
Gaps in the monthly series, age of the last import, cross-checks, single bookings whose total differs from the monthly value.
- 1 rules
Anomalies
Jumps, outliers, unexpected signs, breaks in a series. They raise a question but do not claim an error.
How far is your data complete?
The data coverage page shows the monthly coverage of chart of accounts, budget, actuals and opening balances at a glance, separate from the checks on content. The checks run at import and are stored as a snapshot, so the pages stay fast.
Data coverage 2026
Acknowledged does not mean forgotten.
Every finding can be marked as explained, confirmed or dormant, with person and time. The acknowledgement holds for the data state: when the figures change, the finding comes back. What is broken stays open, and “not checked” looks different from “no findings”.
See your own figures in planist.
25 minutes on video. We show the reconciliation on sample data or on an export from your bookkeeping.