# Bookkeeper · Day 3 — The audit trail your client will actually open

**Subject:** The audit trail your client will actually open
**Send:** Day 3
**Goal:** deepen feature understanding around the audit trail (the
real differentiator vs. spreadsheet workflow)

---

Hi {{first_name}},

Most "data cleaning" tools spit out a clean file and call it done. The thing your *client* needs — and what protects you in a year when they ask "why did you change that?" — is the audit trail.

Here's the file DataTools writes alongside every cleaned export. It's a CSV called `<filename>.audit.csv` and it sits next to the cleaned file in your output folder.

Five columns, append-only:

| original_value | new_value | rule_applied | confidence | timestamp |
|----------------|-----------|--------------|------------|-----------|
| `AMZN Mktp` | `Amazon` | `merchant_canonicalize` | 0.94 | 2026-05-04T09:12:03 |
| `  Starbucks  ` | `Starbucks` | `whitespace_strip` | 1.00 | 2026-05-04T09:12:03 |
| `01/02/26` | `2026-02-01` | `date_normalize_dmy` | 0.88 | 2026-05-04T09:12:03 |

Why this matters in a real client conversation:

- **The client asks "why is this Amazon when my statement says AMZN Mktp?"** — open the audit CSV, point at the `merchant_canonicalize` row. Done in 10 seconds.
- **A reviewer (auditor, accountant, you in 6 months) asks "what changed?"** — the audit CSV is the answer. Diffable, openable in Excel, no proprietary format.
- **You spot a wrong rule firing** — the `confidence` column tells you which rules to tune. Anything <0.90 is worth eyeballing.

One workflow change worth making: when you send the cleaned file to QuickBooks, send the audit CSV to the client at the same time, in a folder labeled "month-end audit trail". Most clients won't open it. The 10% that do will trust you forever.

Reply if you want me to walk through the audit format on a call — happy to do a quick screen-share for any buyer in the first 30 days.

— Michael
{{support_email}}