The Explorer is the ledger as a pivot table that keeps going down: a caption opens its accounts, an account its months, a figure its journal lines, a line its whole entry. Then profile what you are looking at, see where its money moved, and draw the sample — on your machine, on every line.
Rows and columns by caption, group, account, month, journal type, user or statement, in net amount, debits, credits, lines or entries. Click a row to open the next level down, any figure to list its lines, any line to read the whole journal. Each row carries its twelve-month trend; the view lives in the address, so Back undoes a drill.
With last year's ledger loaded, Change lines the two years up with the difference and the percentage, the largest change first, against a band drawn at materiality — and one tick keeps only the changes over it.
Any table exports to Excel. "Keep in working papers" files it with the engagement under a reference, EX-001 onward, and the exported workbook carries it as its own tab.
The questions to ask an account, a caption or a user before testing it, answered on one screen — and each mark opens its entries.
The share of the slice's entries carrying each of the signals journal entry testing asks about:
Lines stratified on the engagement's own thresholds — under clearly trivial, up to half of performance materiality, up to performance materiality, and above — with the share of the value in each band. The basis for choosing what to test, and how.
The Benford first-digit test on the slice, its conformity read on the mean absolute deviation as Nigrini sets the bands, and a digit outside the 99% band shown in amber. Each digit opens its lines.
Month by month against last year, the day of the week, and who posts when — with the users who posted to the slice this year and never last year named as new.
The accounts on the other side of the slice's entries, and across the whole ledger the debit and credit pairings that last year never saw.
A map of the year from the caption credited to the caption debited, each entry prorated across its lines, so the routes the business takes — revenue to receivables, receivables to cash — stand out by their width, and the ones it should not take stand out by being there at all. With last year loaded, a flow the ledger did not have is drawn in amber.
A band opens its entries; a caption opens its profile.
Sample from what is on screen — a caption, an account, a user, a month — on its debits, its credits or both. The method, the side and the seed are written on the sample, and the order is fixed, so a reviewer re-performing it gets the same lines.
The interval is tolerable misstatement ÷ the reliability factor for 80, 90 or 95% confidence; a seeded start, then one unit every interval along the cumulative value. A line worth at least the interval is always selected, and says so.
A set number of lines, ordered by a hash of their key and the seed. The same seed gives the same lines, on any machine.
Every line at or above a threshold — performance materiality unless you set another — and a set number more at random from the rest.
ISA 530 and AU-C 530 sampling, documented: each sample exports to Excel, and kept in the working papers it becomes a tab of the engagement's workbook under its EX reference.
One free engagement to evaluate, valid 60 days. The ledger never leaves the machine.