Believe it or not, there are plenty of similarities between coffee and telecoms services (voice, data). Both are largely commoditised product. Then it shouldn’t come as a surprise that telecoms service providers and Starbucks have so much in common. For a start: discrimination. Price discrimination.

Price discrimination in telecoms (and in coffee shops)

There is plenty of price discrimination around us. People don’t pay the same for essentially the same goods or services. If that sounds unfair, then think of ‘discrimination’ as ‘differentiation’ instead. Economically speaking, price discrimination is often socially optimal as:

  • it allows people who can afford less to have access to products / services at a lower price; otherwise, they would be excluded from the market;
  • it allows businesses to charge more to those who are ready to pay more, and less to those who are only ready to pay less; by serving both customer categories, businesses produce more and maximise their revenues.

So how do you discriminate in practice? This is how it works in telecoms businesses, and in coffee shops.

‘Unique target’ price discrimination

Ideally, as a business, you want to charge as much as individual customers are ready to pay. This is called the ‘unique target’ approach. But finding out how much someone is ready to pay can be very time consuming, so it is mostly applied when the item sold has a high-value compared to the value of the seller’s time. You can find a lot of that in B2B businesses for high-price items.

But also in retail: in the age of the Internet, ‘cookies’ installed on PCs and sophisticated data-mining techniques, customers also leave a lot of hints about what they preferences are and what they buy.  Companies like Amazon used to price their goods based on records of individual customers. When people found out, they were furious. In the meantime, Amazon has stopped this practice.

‘Group target’ discrimination

If the individual target approach is rather unpopular, people don’t seem to mind ‘group targeting’ so much. The idea is the following: there is a special price for children in the bus, senior in the train, students at museum, or locals at the swimming pool. Or there is a special (higher) price for tourists because they are ready to pay more than the locals. Intuitively, this practice seems fair because there is a high correlation between people who can afford less and those who are willing to pay less. And those who can afford more are usually willing to pay more – but not always.

Group targeting is not as effective as individual targeting, but it is easier to put into practice – requiring less information – and also socially more acceptable. In any case it delivers more profits for companies, and it might even put companies in a positive light, showing that they care about those who have less money.

‘Self-incrimination’ approach

But there is yet a better approach to price discrimination, based on letting people signal that they don’t worry so much about price. This is how it works: companies sell the same product in different quantities, or slightly different product at different prices, or the same product in different locations.

Think about Starbucks: you can buy your coffee ‘short’, ‘tall’, ‘grande’ or ‘venti’ (which is giant!). But also

  • “Milk or no milk”?
  • “Sugar, sweetener or no sugar”?
  • “Normal or decaf”?
  • “Blonde, dark, medium or flavoured”?
  • “Caramel, vanilla, pumpkin spice or moka”?
  • “Ristretto, half-caf, stirred or macchiato style”?

So you can order a “Decaf Double Tall Non-fat Extra-dry Cappuccino”, a “Grande Double Chocolaty Chip Frappuccino Blended Creme”, or a “For Here, Iced Half-Caf, Quad, Grande, Soy, Starbucks Doubleshot™ on Ice + Energy”.

Aaarrrgh,  I don’t care, just give me coffee.

The cost of all these coffees is essentially the same (it does not really cost more to make a ‘tall’ coffee rather than a ‘short’ one), but the prices charged can be very different.  From the consumer perspective, you might even be compelled to order a ‘grande’ rather than a ‘tall’ one, because you don’t want to sound ‘cheap’.

In telecoms, service providers use clever pricing strategies to discriminate and, as much as possible, let their customers ‘self-incriminate’. There are plenty of ways to discriminate:

  • by group, e.g. tariffs for people who are below 20 years old only;
  • by location, for example home-zone tariff;
  • by volume, even when the cost of a higher volume is essentially the same as a lower one – as in the coffee example above;
  • by contract type, in particular with prepaid and postpaid tariffs being quite different; or different tariffs for private or corporate users.

Quite often, service providers overdo it. Finding the right tariff for end-users can not only be really confusing, but near impossible without Excel and a Ph.D. – not unlike ordering coffee at Starbucks. Let us take a look at a first example to better understand how service providers do it – or believe they do it.

Service providers use price discrimination to make more money from us….

Our service provider offers a mobile broadband connectivity service for laptop users, with the following three packages: $9 for 1 GB, $19 for 5 GB, or $29 for 10 GB. Clearly the ‘apparent’ price per GB for the ‘large’ package is much lower than for the entry package, which serves as incentive for users to upgrade to a larger package. This is reinforced by the price for additional GB, should the user exceed the allowance in his package.

But is the ‘large’ package really a bargain? It seems attractive for users, but the truth is, it is even more attractive for the service provider. A true ‘win-win’, but the best thing is that although customers believe that they get a discount when they take a higher package, they actually pay more per GB and generate more profits to the service provider. That’s really clever. Let us look at the economics on the service provider side to understand why.

Let us assume that the ‘fixed cost’ (some people call this the ‘cost to serve’) are $7 per sub per month. Let us also assume that the cost per GB is $1.  Then the following picture emerges: ‘Large’ users pay, after deducting the fixed costs, $22 for 10GB, whereas ‘Small’ users pay, after fixed costs, only $2 for 1 GB. It is clear that from the service provider’s point of view, ‘large’ package customers not only consume a lot more, but they also have a higher profit margin per GB!

At least from that perspective, you would say that ‘large’ users are much more attractive that ‘small’ users. The economics above are not unlike the Starbucks economics, where people take all sorts of extras and larger size for a much higher price, but essentially the same cost to Starbucks, so a much higher profit! And the good news is that when customers are on the wrong package i.e. the ‘small’ users consuming 3 GB instead of 1 GB, the service provider makes a huge amount of money on the extra GB.

More money does not always mean more profit

But sometimes, the economics work the other way round, as shown below. If the fixed costs are $5 per month, and the cost per GB $2, then the picture changes.  ‘Large’ package customers are still more profitable in absolute value, simply because they consume a lot more than ‘small’ package users.  However, their profit margin is actually lower. The reason is that after fixed costs, they pay $24 for 10 GB, whereas ‘small’ users pay $4 per GB after fixed costs, so have a higher profit margin on the GB that they consume.


In the telecoms industry, low ARPU customers are not always less profitable. Let us take a look at a second example that emphasises this simple truth.

Most profitable customers are not always those you believe

Another common belief in telecom is that postpaid customers are ‘high-value’, and prepaid customers ‘low-value’. But the truth is, ARPU and profits are not quite the same, and high-ARPU customers are only high-value if they also generate high profits, as shown below.

Let us imagine a service provider with two prepaid and postpaid packages for 60 minutes of voice, priced at $10 and $9 respectively. The lower price for the postpaid package, as well as the apparently lower price per min – $0.15 instead of $0.17 –  act as incentive for users to choose a postpaid contract.

Are postpaid customers ‘great’ for this service provider? Once again, it depends on the service provider cost base. Let us first assume that the economics are as follows.

The ‘Postpaid’ package, under the assumption above, has a lower fixed cost (no scratch card costs, longer life cycle etc).  Its margin and profit margin are much higher than for the ‘Prepaid’ package.

But in some cases, the economics are actually the other way round, with high fixed costs for postpaid users (invoicing, bills inquiries, claim management, high acquisition costs etc).  In that case, the valuable customers are actually the prepaid customers, not the postpaid customers.

Summa summarum

The first lesson is that all customers are not created equal, and price-discrimination is used not only to ‘de-commoditize’ product and services, but also to let users tell us whether they are ready to pay more, or not. Or whether they are clever at finding the right package for them, or not, like the guys taking the $9 package for 1 GB but consuming 3 GB every month – but maybe this has nothing to do with cleverness, they simply tell us that they don’t care how much they pay, and self-incriminate, thereby generating great profits for the operator.

The second lesson is that unfortunately, many service providers are not clear about their cost base and the economics of their various products and services, and often don’t know who their truly profitable customers are. Worse: they believe that they know, when they don’t know that they don’t know. Unlike coffee, those paying more and not always the more profitable customers…