Switching Costs

Switching costs are everything that switching providers costs – not just in dollars. Someone who keeps their phone but switches their mobile contract has to compare offers, cancel a contract, and wait for the new SIM card. This effort is a switching cost, even if no bill arrives for it. On top of that come monetary amounts, such as a cancellation fee or a new device that has to be bought. And then there’s the risk: you don’t know in advance for certain whether the new provider is really better. Experts add all three parts together when they talk about switching costs.

Why loyal customers are worth more than good products

For companies, high switching costs are a shield. A competitor can build a better product and still fail to win customers – because switching is too much hassle. Investor Warren Buffett calls this an economic moat: a barrier that protects the business from attackers. Switching costs are one of the four or five most important moats there are.

This has direct consequences for pricing. If switching means 5,000 dollars of effort, the provider can raise prices by a few percent without losing customers. This is exactly why analysts look at churn rate for software companies. A low churn rate is an indication that customers are effectively locked in.

For customers, the same effect is a disadvantage. In the end, you pay more or use worse technology, just because leaving would be unpleasant. Competition authorities pay attention to this: if a provider artificially makes switching difficult, this can become legally problematic. This is why the EU enforced mobile number portability.

What the hurdle actually consists of

The first building block is your own data. Anyone who has stored photos, documents, or notes with a provider for years has to somehow take that volume of data with them. Often the file formats don’t fit with the new provider, and some of it gets lost. This is called data lock-in.

The second building block is learned knowledge. A graphic designer who has worked with an image editing program for ten years knows every keyboard shortcut. When switching to a different program, they are slower for weeks. This lost work time is a real, measurable cost – for companies, often the biggest of all.

The third building block is connections to other systems. In companies, dozens of other programs are often attached to a central piece of software, exchanging data with it. If you rip out the core piece, every single connection has to be rebuilt. Such projects take years and cost millions. That’s why large companies almost never switch their administrative software voluntarily.

Switching costs among AI providers and in your everyday life

In the AI industry, this is currently a central point of contention. Anyone who builds an app on a particular language model tailors their instructions precisely to its quirks. If they switch providers, many of these instructions no longer work equally well. In addition, cloud providers tie customers in via proprietary chips and tools that are only available from them.

This is why the success of open models, which you can download and run yourself, is so important. They lower switching costs and thereby push down the prices of the big providers. Conversely: the more deeply an AI is embedded into a company’s workflows, the higher the hurdle rises again.

In everyday life, you encounter the same principle with your phone’s operating system. Purchased apps, accessories, and the group chat with friends only apply within one of the two ecosystems. A common mistake is to confuse switching costs with customer satisfaction. Someone who stays because switching is too much of a hassle is not satisfied – just locked in.