Retail Revenue & Customer Concentration

Business context

A retail and distribution business reviewed revenue through standard period reports: totals, growth against prior periods and performance by product. Those views answered “how much?” well. They did not answer “from whom?” or “how exposed are we if a few relationships change?”

Challenge

Aggregated reporting hides concentration. A business can grow overall while becoming more dependent on fewer customers, and nothing in a totals-based report will show it. The challenge was to restructure the historical data so that the distribution of revenue across customers and regions became visible.

My contribution

• Prepared and modelled the historical sales data at customer level.
• Ranked customers by revenue contribution and built cumulative-share views.
• Segmented the results by region, product group and period.
• Presented the finding and its commercial implications in a form leadership could act on.

Investigation and methodology

The core technique was a Pareto-style analysis: customers sorted by revenue, with cumulative share of revenue plotted against cumulative share of customers. Segmenting the same curve by region showed whether concentration was uniform or localised. It was localised — the high-value segment sat mostly in one region.

Solution design and implementation

The analysis was delivered as a Power BI report, used by senior executives, with filters for region, customer segment, product group and period, so the concentration could be examined from several angles rather than as a single static number.

Results and business value

In the historical dataset examined, approximately 5% of customers accounted for approximately 95% of revenue, and that segment was concentrated in one region. The value of the finding is strategic: it identifies a dependency risk, informs where account management attention belongs, and gives a baseline for tracking whether revenue becomes more or less diversified over time.

Technical details

• Ranking: customers ordered by revenue within the selected filters; cumulative revenue share calculated over the ordered list.
• Concentration measures: share of revenue held by the top 5%, 10% and 20% of customers; regional split of the top segment.
• Model: customer, region, product and date dimensions related to a sales fact table, so every filter recalculates the ranking rather than reusing a fixed list.
• Headline figure: approximately 95% of revenue from approximately 5% of customers (rounded; historical dataset only).

Lessons, constraints and improvement opportunities

• Concentration is a property of the distribution, not the total; it has to be designed into reporting deliberately.
• The analysis describes the past. Its value depends on whether it is tracked over time and connected to account decisions.
• A useful extension is a concentration index monitored each period, with alerts if dependency on the top segment increases.