Cavo Yachting · Piraeus
Cavo Yachting: which customers are worth the most, and which are about to leave
Sales were known. What could not be seen was which customers bring the value, which buy again, and which have started to go quiet.
- Sector
- Boat equipment trade
- Services
- Sales analysis · Customer value and groups · Business intelligence
- Tools
- Power BI · DAX · Python
The question
Cavo Yachting trades in boat equipment. Its sales were known by month and by product. Management’s questions, though, were about customers:
- Which customers bring most of the value?
- How many of those who buy once come back?
- Which good customers have started to go quiet, before they are lost?
The answers were inside the company’s sales and purchasing data. They had just never been organised around the customer.
What we built
- A data model around the customer. Four years of sales and purchases, organised with shared tables for customers, products, channels, salespeople and calendar.
- A sales board. Sales by year, product category, customer, channel and salesperson, compared with the previous year.
- Customer lifetime value (CLV). What a customer leaves over the whole relationship with the company, not only in their latest order.
- Customer groups (RFM). Every customer is ranked by how recently they bought, how often, and how much value they brought. That separates the top customers, the loyal ones and those at risk of being lost.
- Repeat purchases (cohort analysis). For each year of first purchase, how many customers bought again the next year and the one after.
- Clustering in Python. A check of the groups with statistical clustering, beyond the RFM rules.
Along with the tools came a written report for management, in question-and-answer form: each page starts from a business question and ends with what the data shows.
What changes in practice
The company no longer sees only how much it sold. It sees to whom.
The customer groups point to specific moves: who the sales team should look after first, who to talk to again before they leave, and where time is not worth spending. The tool is set up to refresh with new data.
The next steps proposed were sales forecasting, product recommendations per customer, and alerts when a customer’s sales drop.
Why it applies to any trading business
Almost every business that sells to repeat customers already has the data for this analysis: its invoices. The same goes for a hotel, with the guests who come back.
The screen at the top of the page is a sample, with made-up data. It does not show Cavo Yachting’s figures or reports.