Spot Club Members About to Cancel With One Export and Claude Cowork

You are my wine club retention analyst. I run [WINERY / BREWERY / DISTILLERY NAME], a [size, for example 900 member] club in [REGION]. Our tiers are [LIST TIERS AND PRICE PER SHIPMENT]. We ship [NUMBER] times a year. Our average member lifetime value is about [$AMOUNT] and our biggest retention worry is [for example members who joined in the tasting room and never came back].

In the folder I pointed you at you will find two exports: a member list and an order history. Read both, match them on member ID or email, and do the following.

1. Tell me first what you can and cannot measure with these files. If a column I would need is missing, name it and tell me what to export next time. Do not guess at data that is not there.

2. Analyze the members who already cancelled. Look at their behavior in the 6 to 12 months before they left: shipment skips, falling order value, gaps between purchases, declined cards, no events, no add-on orders. List the patterns you find, most predictive first, with the number of former members showing each one.

3. Apply those patterns to active members. Sort them into three groups: High Risk, Watch, and Healthy. Give me the count in each group and the total annual revenue sitting in High Risk and Watch.

4. For High Risk and Watch, produce a table with member name, email, tier, join date, sign-up source, months as a member, the specific signals that flagged them, and a one line note on the likely reason.

5. Break High Risk into no more than 4 named sub-groups based on why they are at risk (for example "declined card never recovered" or "tasting room signup, never ordered again").

6. For each sub-group, write a personal save email from [YOUR NAME, TITLE]. Under 150 words. Plain, warm, first person, no marketing voice. Lead with a specific reason we are reaching out, offer one concrete fix (pause, tier change, swap the reds for whites, a call, an update to their card), and ask one question. Do not offer a discount unless I tell you to. Include a subject line and a merge field for the first name.

7. Give me a one page action plan: who to email this week, who should get a phone call instead, and what to track 30 days out to see if it worked.

Save everything as a Word document and the at-risk table as a spreadsheet in the same folder. Push back if my assumptions about why people leave do not match what the data says.