TIKTOK SHOP · JUN 10, 2026 · 6 MIN

How to Automate TikTok Shop Weekly Reporting with Claude and the Cruva MCP

A weekly TikTok Shop affiliate report used to be an afternoon of dashboard exports and spreadsheet assembly. With Claude connected to the Cruva MCP, it is a single prompt and under two minutes. Here is the exact setup, the prompt structure, and the report format clients actually read.

Weekly client reporting is the least leveraged hour in a TikTok Shop affiliate program. The data already exists in your affiliate platform. The format is the same every week. The only thing that changes is the date range. That is exactly the shape of work an MCP connector is built to remove.

Here is the short version: connect Claude to the Cruva MCP, give it one structured prompt with your exact date range, and it returns a client-ready Excel report with week-over-week deltas, top affiliate rankings, top content, and prioritized insights. End to end, the run takes less than two minutes.

This piece walks through the exact setup, the prompt structure we use for weekly affiliate reports, and where teams get it wrong.

What You Need Before You Start

Three things, none of which require engineering time:

How to Run the Report, Step by Step

Step 1: Turn On the Connector

Open Claude, hit the plus button, and check that your connectors are connected. Toggle on Cruva (or whichever platform holds your affiliate data). This is what allows Claude to call your tools and fetch live program data rather than working from whatever you paste in.

Step 2: Give It the Structured Prompt

Do not ask for "a weekly report." Ask for the specific report your client expects, in the order they read it. The structure that has worked for us is four blocks, covered in detail in the next section: program overview with week-over-week deltas, top affiliates, top content, and prioritized insights.

Step 3: Set the Exact Dates

Put your date range directly in the prompt, explicitly. "Last week" is ambiguous. "June 1 through June 7, 2026, compared against May 25 through May 31, 2026" is not. If the dates are vague, the report fetches the wrong time range, and every delta in it is wrong for your client.

Step 4: Run It and Open the Excel File

Claude returns an Excel file. Open it directly in the chat. The deltas are computed, declining metrics are highlighted in red, and everything is categorized into the sections you asked for.

Step 5: Move It Into Your Client Reporting

From there it is a copy-paste into the master spreadsheet alongside the rest of your client reports. Or, once Google Drive is authorized as a connector, save it straight to Drive and skip the copy step entirely.

The Prompt Structure That Produces a Client-Ready Report

The report has four blocks, and the order matters because it mirrors how a client actually reads it: how is the program trending, who is driving it, what content is working, and what are we doing next.

Block 1: The program overview. Week-over-week deltas for every headline metric: GMV, video views, first-time posters, live sessions, and all outreach activity. This is the quantitative case for what has been working, and it is where new tests show up as movement in the numbers rather than as claims in a status email.

Block 2: The top affiliates. The top 10 creators in the program, ranked by GMV, with units sold and conversion alongside. This tells the client where revenue is concentrating and which creators are worth deepening.

Block 3: The top content. The top five videos for the brand and the top live streams for the week. This is the creative signal: what hooks, formats, and products the audience is actually responding to.

Block 4: The key insights. Organized by priority, high, medium, and low, and color coded. The goal is that a client can scan the data and immediately understand where you are taking action next, without a call to interpret it.

A condensed version of the prompt, ready to adapt:

Pull my TikTok Shop affiliate data from Cruva for [DATE RANGE], compared against [PRIOR WEEK RANGE]. Build a weekly report with: (1) week-over-week deltas for GMV, video views, first-time posters, live sessions, and outreach, (2) top 10 affiliates ranked by GMV with units sold and conversion, (3) top 5 videos and top live streams, (4) key insights prioritized high, medium, low and color coded by priority. Return it as an Excel file. Highlight declining metrics in red.

Fill in the brackets with real dates every single week. The structure stays fixed; the dates are the only variable.

Why This Beats the Manual Version

The manual version of this report is a tour through dashboards: pull GMV here, export creator stats there, screenshot the top videos, compute the deltas by hand, and assemble it all into a deck or spreadsheet. Nothing about that work is judgment. It is collation, and collation is what the connector removes.

What stays human is the part that was always the actual job: reading the insights block, deciding which actions to take, and framing the story for the client. The report automation does not replace the account manager. It returns the hour the account manager was spending as a data clerk.

It also makes the reporting more consistent. The same prompt produces the same structure every week, across every client, which means week-over-week comparisons are real comparisons and a new account manager inherits a format instead of a habit.

Where Teams Get This Wrong

Vague date ranges. The model will fetch something for "last week." Whether it is your client's reporting week is another matter. Exact dates, in the prompt, every time.

Connector not actually enabled. If Cruva is not toggled on, you get a plausible-looking report built from nothing. Check the connector before you run, and spot-check a number you know against the platform the first few times.

Treating the output as the deliverable. The Excel file is the input to your client communication, not the communication itself. The insights block is a draft of your point of view, and the clients are paying for the point of view.

Where This Fits in the Broader Affiliate Stack

Reporting is the read side of an automated affiliate program. The write side is the same connector pattern pointed at action: AI agents handling creator outreach and AI-generated briefs that scale creator output without losing brand control. The teams getting the most out of MCP connectors are running both sides off the same data layer.

If you want help standing up the reporting workflow, or the rest of the affiliate operations stack around it, clankersapp.com is where to start.

Frequently Asked Questions

What is the Cruva MCP? Cruva is a TikTok Shop affiliate management platform, and its MCP (Model Context Protocol) connector lets an AI assistant like Claude pull your shop and affiliate data directly. Once the connector is enabled, Claude can fetch GMV, video, live session, and outreach metrics on request, without you exporting anything by hand.

Do I need to write code to automate TikTok Shop reporting this way? No. The setup is toggling on a connector inside Claude and writing a structured prompt in plain English. There is no API integration to build and no script to maintain.

What metrics belong in a weekly TikTok Shop affiliate report? Week-over-week deltas for GMV, video views, first-time posters, live sessions, and outreach activity, plus top affiliates ranked by GMV, units sold, and conversion, top videos and live streams, and a prioritized, color-coded list of key insights.

Can the report be saved directly to Google Drive? Yes. The workflow returns an Excel file you can open in the chat and copy into your client reporting spreadsheet, or, once Google Drive is authorized, save directly to Drive.

Does this work with outreach tools other than Cruva? Yes. Any tool that exposes an MCP connector plugs into the same workflow. Reacher and Yuka are two others in the TikTok Shop affiliate space that connect the same way.