What to include in a client data delivery report (template inside)

Two weeks after you delivered the cleaned dataset, the client forwards an email: "My director asked in the Monday meeting where the 18% churn figure came from, and I didn't know what to tell her." You sent a clean file and a one-line "here you go" a fortnight ago, and now neither of you can reconstruct which rows you dropped, how you defined churn, or why the number moved from the client's own estimate. The data was fine. What was missing was the one-page report that would have let the client answer her director without picking up the phone to you. That report is the difference between a file and a deliverable.

Why the file alone is not the deliverable

A cleaned dataset with no explanation puts the client in an impossible spot: they have to defend numbers they didn't produce and can't see the working behind. The first time someone senior questions a figure, the client's trust in the whole dataset wobbles, and it lands back on you as a "can you double-check this?" that costs an hour and some credibility. A delivery report pre-empts all of it. It travels with the data, answers the predictable questions before they're asked, and — quietly — documents that the cleaning was real work worth paying for.

The template, filled in

Here's the whole thing on one page. Copy the structure; swap in your numbers.

DATA DELIVERY REPORT
Client: Northbridge Retail   Dataset: Q2 churn analysis
Delivered: 7 July 2026       Prepared by: [you]

1. SOURCE
   orders_q2.csv (4,812 rows) + subscriptions.csv (1,190 rows)
   Received 1 July 2026 from client finance team

2. WHAT WAS CLEANED
   - 2,300 order dates converted from text to real dates
   - 14 duplicate orders removed (CRM audit rows)
   - Region mapped to 8 approved values (see mapping tab)
   - 37 rows with no Order ID excluded (listed in 'excluded' tab)

3. KEY DEFINITIONS
   - Churn = subscriptions active in Q1 but not renewed in Q2
   - Revenue = net of refunds, GBP

4. RECONCILIATION
   Total revenue £474,300 — matches client finance summary ✓
   Active subscribers 1,153 — client CRM says 1,150 (3 pending)

5. KNOWN LIMITATIONS
   - 37 excluded rows (0.8%) had no Order ID; not in totals
   - June data is provisional per client note

6. FILES DELIVERED
   q2_clean.csv, mapping.xlsx, excluded_rows.csv

Notice what section 4 does. When the director asks about churn, the client opens the report, sees the definition and the reconciliation, and answers in the meeting — no email to you, no wobble. The 18% is now a number with a paper trail instead of a figure that appeared from nowhere.

The six sections, and why each earns its place

1. Source — which files, how many rows, received when. Anchors everything to a specific input you can point back to.
2. What was cleaned — the plain-language change log. This is the section that makes your invisible work visible.
3. Key definitions — how "churn," "active," "revenue" are defined here, so no two people read the same word differently.
4. Reconciliation — your totals versus a number the client already trusts. The single most reassuring line in the document.
5. Known limitations — excluded rows, provisional data, anything you didn't fix. Naming a limitation builds more trust than hiding it.
6. Files delivered — an inventory, so nothing gets lost and the client knows what they have.

Six short sections, fifteen minutes to fill in, and it converts "here's the file" into "here's the file, and here's exactly how to trust it."

Send it every time, even when nobody asked

The temptation is to skip the report on small jobs or friendly clients, but those are exactly the projects that come back to bite you six weeks later when someone senior gets involved. Make it a fixed part of every delivery, the way you'd never send code without a commit message. It costs you a quarter-hour, it saves the client a meeting, and it puts your name on a document that says, in writing, that the numbers were checked — which is worth far more than the file itself.