Measuring Dependencies: Customers, Suppliers, Employees in One Number

Why dependence needs a number

Dependence is rarely a secret in small businesses. The owner knows that the largest customer carries a lot of revenue, that one component comes from only one manufacturer, and that bookkeeping depends on one colleague. As long as this knowledge remains a feeling, it can neither be compared nor tracked over the years. One risk seems more threatening than another because it caused trouble recently, not because it is larger.

This guide therefore answers one question: How do you measure dependence on customers, suppliers, and employees so that they become comparable and you know where to counteract first? To answer it, we first clarify the terms and the mental model, then introduce the metric, use a study to examine when concentration hurts and when it does not, and finally work through an example you can apply to your own business.

What you do not measure, you either fear too much or not at all.

What dependence means: the resource dependence model

In 1978, Jeffrey Pfeffer and Gerald Salancik presented resource dependence theory. Their starting point: no company is self-sufficient. It needs orders, materials, money, and knowledge from outside, and whoever controls these resources has power over the company. According to the theory, the size of that power depends on two quantities.

• Importance: How much of the business depends on this one source? A customer with a three percent share of revenue is insignificant, one with thirty percent is not.
• Replaceability: Are there alternatives, and how quickly could they be developed? A supplier with a large share is harmless if two equivalent vendors could step in at any time.

Dependence only arises when both come together. An everyday analogy makes this tangible: if you have only one key to your front door, you are not at risk because there is only one, but because the locksmith takes three hours on a Sunday. The theory also says that dependence is rarely one-sided. A large customer who would find it hard to replace their supplier is dependent too. What matters is the ratio of the two dependencies, that is, the power imbalance.

The metric: Herfindahl index and effective number

For importance, a measure from competition economics is enough. The common measure of concentration is the Herfindahl-Hirschman Index (HHI). You take the market shares of all providers in percent, square them, and add up the results. A single provider with 100 percent yields 10,000 points, many small providers yield a value close to zero. This is also how the analysis when a few providers dominate a market describes it. Instead of market shares, you insert the shares of your customers in revenue, of your suppliers in purchasing, or of your employees in a critical task.

The point score is abstract. It becomes easier to understand when you divide 10,000 by the index. The result is called the effective number: it tells you how many equally sized partners your actual distribution corresponds to. A business with 40 customers and an index of 1,250 effectively has eight equally sized customers. The other 32 create work but carry hardly any of the risk.

A word on interpreting the values: since their 2023 merger guidelines, the U.S. competition authorities have classified markets above 1,800 points as highly concentrated. For a customer base, this is no benchmark, because a market and a customer base are different things. I therefore use the effective number and a rule of thumb of my own, given further below.

A worked example: three dependencies compared

An example business with 1.2 million euros in revenue has 35 customers. The largest accounts for 30 percent, the next four for 15, 10, 8, and 7 percent, and the remaining 30 customers for 1 percent each. In purchasing, 60 percent comes from one supplier and 20 percent each from two others. The owner manages 70 percent of customer relationships, a sales employee 20 percent, and a project manager 10 percent.

The result surprises many owners. The large customer with 30 percent feels like the biggest risk, and it is indeed above the rule of thumb from the analysis customer concentration risk, according to which you should deliberately counteract once a single customer accounts for a quarter of revenue. Measured by concentration, however, customer management is the tightest point: the business behaves as if it had barely two salespeople, and one of them is the owner. If the owner drops out, 70 percent of relationships are shaken, including the one with the large customer.

Is concentration always bad? Two perspectives

The first perspective regards any concentration as a risk to be reduced. It is supported by the study by Saboo, Kumar, and Anand (2017), which found lower profitability at high customer concentration because bargaining power shifts to the customer. The second perspective sees concentration as a strength: with few partners you know them better, work more closely with them, and are more likely to get help in an emergency. There is evidence for this too, such as the inverted U-shaped relationship in Han, Yu, Li, and colleagues (2023), in which a moderate concentration on large customers went along with the best financial performance.

Jiang, Yeung, Han, and Huo (2023) tested both perspectives in a crisis. They analyzed 23,440 quarterly observations of 2,990 listed Chinese companies from 2019 and 2020 and also interviewed 24 managers. They measured concentration as the share of the five largest customers in sales and of the five largest suppliers in purchasing, and resilience as the change in productivity. In the disruption stage, when demand fell, high customer concentration hurt, while high supplier concentration helped. In the recovery stage, when materials became scarce, the supplier effect reversed: now concentration hurt, while customer concentration no longer had any effect.

The authors explain this with resource dependence: concentration becomes dangerous when the power imbalance worsens to the business’s disadvantage. I consider this a more important finding than the question of whether concentration is good or bad. It means the same number means different things in different situations. The study’s limits are clear: it concerns large listed companies in China and a particular crisis. The direction of the mechanism, however, carries over to a small business that bundles its purchasing with one supplier.

The dependence traffic light: from number to decision

So that the number leads to decisions, I combine it with replaceability. For each of the three areas, you go through the same four steps.

1. Record the shares

   List the shares of customers, suppliers, or people for the last fiscal year. You can combine small positions below one percent; they barely change the result.

2. Calculate index and effective number

   Square the percentages, add them up, and divide 10,000 by the result. A spreadsheet does this in one row.

3. Check replaceability

   For customers with the Replacement-Time Test: revenue of the largest customer divided by new annual revenue won per month. For people with the bus factor from the guide organizing deputies. For suppliers with the question of how long approving a second vendor would take.

4. Set the traffic light

   Classify each area according to the rule of thumb below and choose one measure for next quarter for the area with the worst color.

In the example business, this gives: customers yellow, because the large customer is above a quarter and the replacement time is long. Suppliers yellow or red, depending on how quickly a second vendor can be approved. Account management red. The first measure would therefore not be acquiring new customers but handing over customer relationships to a second person. How to do that is described in the essay succession and sales. On the supplier side, the review in the guide evaluating suppliers helps.

What the metric does not do

One number for all three areas has an advantage: it shows where the business is tightest, even where nobody is looking. But it has limits worth knowing. First, it measures only the distribution, not the quality of the relationship. A customer with a long contract and several contacts is less risky than one who can cancel monthly, despite the same share. Second, it does not see hidden connections. If three customers belong to the same industry, they may drop out together in an industry crisis. How to account for that is described in the guide balancing the customer portfolio.

Third, the meaning depends on timing, as the study by Jiang and colleagues shows. Concentrated purchasing can help in a demand slump and hurt during material shortages. The metric does not replace judgment; it directs it. For the business’s risk register, as described in the guide risk management for small businesses, it is nevertheless a good basis, because it can be calculated the same way every year.

Conclusion and a first step

To the guiding question: dependence on customers, suppliers, and employees can be made comparable with the same metric if you square the shares, add them up, and convert the result into the effective number. Resource dependence theory reminds us that this is only half the answer. Concentration only becomes dangerous when the partner is hard to replace and the power imbalance tips against you. That is why the number always comes with a check of replaceability.

1. Export revenue per customer and purchases per supplier for the last fiscal year.
2. For account management, estimate which share of revenue sits with which person.
3. Calculate the index and the effective number for all three areas.
4. Check replaceability and set the traffic light.
5. For the area with the worst color, choose one measure and a date next quarter.

The sober observation is this: the most dangerous cluster in a business is often not the largest customer but the person who manages that customer.