Every business eventually runs into the same problem. Getting a customer to sign up once is hard enough, but getting them to stay is a different game entirely, and it rarely has anything to do with how good the product is on any given day. The companies that hold onto customers for years, sometimes decades, have usually built some form of lock-in into the relationship without the customer fully noticing it happen.
Economists Carl Shapiro and Hal Varian mapped this out in their book Information Rules, where they laid out nine distinct types of lock-in that businesses rely on to keep customers from walking away. Each one works through a different kind of switching cost, and most successful companies end up combining several of these without ever calling it a “retention strategy” out loud. Here’s a breakdown of all nine, and how they actually show up in the real world.
Brand Loyalty
This is the oldest and most emotional form of lock-in. A customer sticks with a brand not because switching is expensive or inconvenient, but because they’ve built a psychological attachment to it. Apple users are the textbook example here. Moving from an iPhone to an Android phone is technically simple, but for many users it feels like giving something up, and there’s a real fear that whatever comes next just won’t feel the same.
Brand loyalty is the hardest lock-in to manufacture deliberately because it depends on years of consistent experience and storytelling. Companies that get this right tend to talk about their products as identities rather than tools, which is why so much marketing budget at the top end goes into feeling rather than features.
Compatibility
Compatibility lock-in shows up the moment a customer’s existing setup depends on a particular product working smoothly with everything else they already use. Switching away means risking that the new system won’t talk to the old one, and the integration costs of fixing that can be steep.
This is a big reason enterprise software is so notoriously hard to displace. A company running its finance stack on a particular ERP system isn’t just choosing software, it’s choosing every other tool that has to plug into it. Microsoft built an empire partly on this principle, with Office, Windows, and Azure designed to work better together than with anything outside the ecosystem.
Contractual
Sometimes the lock-in is written into the fine print. Contractual lock-in uses early termination fees, multi-year agreements, or the threat of losing existing discounts to make leaving financially painful. Telecom and broadband providers run on this almost entirely. The much-discussed SaaS pricing collapse that Naval Ravikant flagged recently is partly a story about contractual lock-in losing its grip as AI tools make rebuilding software trivially easy, removing one of the oldest forms of customer retention almost overnight.
Gym memberships, enterprise cloud contracts, and insurance policies all lean on this same mechanic. The product doesn’t need to be exceptional if the cost of leaving is high enough on its own.
Loyalty Programs
Airlines perfected this one. A customer who has spent two years accumulating miles toward a free flight is not going to switch carriers easily, even if a competitor offers a cheaper fare tomorrow. The switching cost here isn’t financial in the traditional sense, it’s the loss of accumulated rewards plus the hassle of starting from zero somewhere new.
Credit card companies, hotel chains, and increasingly e-commerce platforms all run versions of this. Amazon Prime functions partly as a loyalty lock-in too, bundling enough perceived value into a single subscription that leaving feels like forfeiting something already paid for.
Data
Data lock-in is quietly one of the most powerful forms on this list, and it’s grown sharper as more of daily life moves online. A customer who has years of photos on one cloud service, or a decade of financial records in one accounting tool, faces real friction in moving that data elsewhere, both in the technical effort of transferring it and the risk of something getting lost or corrupted in the process.
This is increasingly central to debates around AI platforms too. As discussed in a recent piece on Anthropic’s distribution moat, AI labs are trying to build switching costs through integrations, compliance certifications, and enterprise relationships precisely because the underlying models themselves are easy to copy. Data, and the workflows built around it, end up being one of the few durable moats left.
Geographical
This type of lock-in is rooted in physical location rather than digital habit. A customer who lives near one bank branch, one preferred grocery store, or one gym isn’t going to drive across town for a marginally better alternative. The switching cost here is literal, the time and effort of changing where you physically go.
Local businesses have leaned on this for decades, but it’s weakening fast for categories where delivery and remote service have removed the need to be physically present. Banking is a good example of a category where this lock-in used to be strong and has eroded as digital-first banks removed the need for a nearby branch entirely.
Learning Curve
Once a person has invested time learning how a tool works, the idea of starting over with something new becomes its own deterrent. This applies whether it’s a piece of enterprise software, a video editing suite, or even a keyboard shortcut layout someone has built muscle memory around. The switching cost is the temporary drop in productivity while relearning a new system, plus the training time required to get back to where they were.
Adobe’s Creative Suite benefits heavily from this. Designers who’ve spent years inside Photoshop or Illustrator rarely jump to a competitor purely on price, because the time lost relearning workflows usually outweighs the savings.
Network Effects
Network effects lock-in is about who else is already using the product. The value of a platform increases as more people join it, which means leaving doesn’t just cost the individual user something, it costs them access to everyone else still on the platform.
WhatsApp’s dominance in India is one of the cleanest examples of this in action. A look at why Hike never managed to dent WhatsApp’s lead despite offering a more feature-rich product found that network effects alone kept users locked in, since switching meant losing access to the contacts and groups that made the app useful in the first place. The product almost stops mattering once enough people are already there.
Search Costs
The last form of lock-in is the simplest, and often the most underestimated. Customers stay with what they have because finding and evaluating alternatives takes time, money, and effort they’d rather not spend. This is especially true in categories with dozens of similar-looking options, where the perceived gain from switching rarely feels worth the research required to find out if it’s actually better.
Streaming platforms benefit from this constantly. A subscriber sticks with one service not necessarily because it’s the best, but because comparing every alternative’s catalog, pricing, and quality takes more energy than just continuing to pay for what’s already familiar.
What ties all nine together is that none of them require the product to be the best available option. They simply need to make leaving feel more costly, more inconvenient, or more uncertain than staying. Most businesses that retain customers well aren’t relying on just one of these, they’re stacking two or three at once, often without naming the strategy out loud.