Balancing the Customer Portfolio: Industries, Sizes, Regions

Many customers, one risk

A supplier with 60 customers looks broadly positioned. But if 45 of them are automotive suppliers, a crisis in the car industry hits it almost as hard as if it had only one customer. The number of customers says little as long as you do not know how similar they are. The analysis customer concentration risk deals with dependence on individual large customers. This guide is about dependence on groups that move together.

This guide answers one question: How does a business check whether its customer portfolio is too one-sided by industry, size, and region, and how does it balance it without losing its strength? To answer it, we first clarify the term, introduce portfolio theory as the mental model, examine a study that applies it to customers, weigh spreading against focus, and finally go through an analysis in five steps.

Sixty customers from one industry are one customer with sixty billing addresses.

What a customer portfolio is

By the customer portfolio I mean the totality of a business’s customers, viewed by their composition. The ABC customer analysis ranks them by their contribution to the business. Forming segments, as described in the guide building customer segments, groups them by similar needs. The portfolio view asks a third question: how strongly do the revenues of different customer groups depend on each other?

Three characteristics are particularly revealing. Industry determines whether customers are exposed to the same economic cycles, rules, and technology shifts. Size determines payment behavior, bargaining power, and buying style. Region determines whether local events, exchange rates, or political decisions hit several customers at once.

The mental model: Markowitz’s portfolio theory

In 1952, Harry Markowitz described how investors should put together a securities portfolio. His central idea: the risk of a portfolio depends not only on the risk of the individual securities but on how strongly they move together. Two volatile stocks whose prices move in opposite directions add up to a calm portfolio. Two stocks that always rise and fall at the same time add up to double the risk.

The same applies to customers. Tarasi and colleagues (2011) explain that stocks from different industries, countries, and company sizes react to the environment and the economy in their own ways. These are exactly the three characteristics by which a business should check its customers. A customer is the more valuable to the portfolio the less their revenue moves in lockstep with the rest.

An everyday analogy shows the idea. An ice cream seller who earns well in summer does not become safer in winter by offering more flavors. He becomes safer by selling roasted chestnuts in winter. More of the same is not spreading.

What a study with customer data shows

Tarasi, Bolton, Hutt, and Walker (2011) applied portfolio theory to the customers of a large B2B company. They used a seven-year data series and proceeded in four steps: they measured the variability in the customer portfolio, examined which market segments developed similarly or in opposite directions, calculated the weights of the segments in an optimized portfolio, and determined how much return individual customers or segments bring relative to their variability.

They compared the efficient portfolio calculated in this way with the actual customer structure and with a portfolio geared only to maximum profit. In back-testing and forward-testing, the efficient portfolio consistently had lower variability than both. The authors consider the approach particularly suitable for B2B companies serving segments from different sectors of the economy. They recommend adding the risk view as an extra layer over existing customer management practices, not as a replacement.

The limits are clear. The study is based on a single large company. A small business rarely has seven years of clean segment data and cannot reshuffle its portfolio at will, because customers are not stocks you buy overnight. What carries over is the way of thinking: besides a customer’s revenue and contribution margin, what counts is how they behave relative to the others.

A worked example: three portfolios, same number of customers

An example business serves three industries. Revenue with customers from automotive supply fluctuates by about 20 percent from year to year, with customers from healthcare by 5 percent, and with public clients by 3 percent. For simplicity we assume the three industries move independently of one another. Depending on the mix, total revenue fluctuates very differently.

The number of customers is the same in all three cases; the difference lies solely in the mix. With the first mix, the business has to build reserves for a year with 16 percent less revenue, with the third only for 6 percent. That changes how much money it has free for investment and how calm it can stay in price negotiations. Whether the healthcare customers also bring the same contribution margin is a second question the example deliberately leaves open.

Spreading or focus? Two perspectives

The first perspective follows portfolio theory: a business should spread its customers across industries, sizes, and regions so that no single event hits it too hard. The study by Tarasi and colleagues supports this, as does the experience of many suppliers who lost their most important customer group all at once in an industry crisis.

The second perspective warns against spreading for its own sake. A business that does a little work for every industry is the first choice for none. The analysis specialization instead of a general store shows that a narrow niche can bring growth and margin. Every new segment requires expertise, references, and its own approach, and these costs are missing from the portfolio calculation.

I consider both perspectives correct and see the way out in adjacent segments. These are customer groups that need the same competence but follow different cycles. A business that builds control technology for automotive suppliers can sell the same technology to makers of medical devices or to the food industry. It remains a specialist in its craft and at the same time spreads industry risk. The focus then lies on the capability, not on the customer group.

The portfolio analysis in five steps

Revenue data for the last three to five years and one afternoon are enough for the analysis.

1. Assign customers

   Assign each customer an industry, a size class, and a region. Rough classes are enough, such as five to eight industries, three size classes, and three regions.

2. Calculate shares

   For each characteristic, calculate each class’s share of revenue and contribution margin. With the Herfindahl index from the guide measuring dependencies, you can also determine the effective number of industries.

3. Check co-movement

   Look at how the revenues of the classes developed in the last weak years. If they fell together, they offer no balance.

4. Set a target

   For the most one-sided characteristic, set a target share, for example: no industry above 40 percent of revenue in three years. The rule of thumb that no single customer should carry more than a quarter still applies alongside.

5. Choose adjacent segments

   Look for two segments that need your competence and follow different cycles. For them, develop a target customer profile as described in the guide defining target customers.

• No characteristic has a class with more than half the contribution margin.
• At least one customer group developed differently from the largest in the last crisis.
• No single customer carries more than a quarter of revenue.
• For each new segment there is at least one reference or pilot customer.

Conclusion: what a balanced portfolio looks like

To the guiding question: a business checks its customer portfolio by sorting its customers by industry, size, and region, calculating the shares, and looking at which groups moved together in bad times. Markowitz’s portfolio theory explains why this co-movement matters more than the number of customers. The study by Tarasi and colleagues shows with real customer data that a deliberately mixed structure lowers fluctuations.

Balancing works best through adjacent segments that need the same competence but follow different cycles. That way specialization is preserved and risk is spread. For this week, one step is enough: assign your 20 largest customers to an industry and check whether one industry carries more than half your contribution margin.

The sober observation is this: a portfolio is not balanced when it has many customers, but when not all of them have bad years at the same time.