Two products can have identical switching costs and only one of them is defensible. That sounds like a contradiction until you look at where the cost comes from. One product is expensive to leave because the customer would lose years of accumulated value. The other is expensive to leave because someone deliberately buried the export button. Customers can feel the difference, competitors can attack the difference, and increasingly, regulators can legislate the difference away.
Most product strategy decks treat “switching costs” as a single line item on the moat. It is not. It is two very different bets, and product managers keep confusing the fragile one for the durable one.
The moat everyone quotes, and the half they skip
Hamilton Helmer’s 7 Powers put switching costs on the map for a generation of product leaders. In his framework, a switching cost exists when a customer would lose value by moving to an alternative supplier for a future purchase. Because the incumbent already captured that value, a rival has to compensate the customer just to reach parity, which lets the incumbent hold higher prices without losing the account. That is a real, well-defined source of power, and Helmer is careful about it in his conversation on business strategy.
The part teams skip is why the value would be lost. Helmer’s own examples split cleanly into two categories. SAP becoming the operational backbone of a company is one kind: the switching cost is the years of business logic, integrations, and trained staff that would have to be rebuilt. Software vendors giving universities free licenses so graduates arrive already fluent in the tool is another kind: the switching cost is learned skill and habit. Both raise the price of leaving. Neither depends on trapping anyone.
Then there is the third kind that never makes it into the framework because it is not a power at all: making the data hard to get out. That one shows up in the strategy deck under the same heading, and it behaves nothing like the other two.
Friction-based lock-in has a shelf life
Call the fragile version lock-in: the customer stays because leaving is painful, not because staying is valuable. Proprietary export formats, contractual auto-renewal traps, integrations that only run one direction, data you technically own but cannot practically retrieve. It works, right up until it doesn’t, and it fails in three predictable ways.
Competitors buy their way through it. When the switching cost is a fixed dollar amount of pain, a well-funded rival simply absorbs it: free migration services, “we’ll pay your termination fee,” white-glove data import. If the only thing keeping a customer is a one-time cost of leaving, that cost has a price, and someone will pay it to take the account.
Customers resent it, and resentment compounds. A customer who stays because they are trapped is not a retained customer; they are a churned customer waiting for an opening. The distinction between lock-in and genuine stickiness is exactly this. As one analysis of SaaS retention frames it, stickiness means the customer does not want to leave, while lock-in means they cannot, and the second group churns the moment the value equation or the exit cost changes.
And now regulators are changing the exit cost by force. The EU Data Act came into force on September 12, 2025, and it targets friction-based lock-in directly. Cloud and SaaS providers must remove contractual and technical barriers to switching, allow customers to terminate with a maximum two-month notice, and support secure data transfer to a competitor or to the customer’s own infrastructure. Switching charges get reduced immediately and, per Latham & Watkins’ read of the switching requirements, must be eliminated entirely by January 12, 2027. If your moat was the export button, the law just moved it.
Durable switching costs make leaving mean losing
The durable version is the opposite: the customer stays because the product is worth more to them today than it was the day they signed, and most of that added value walks out the door if they leave.
Adobe is the textbook case, and the numbers are the reason. Creative Cloud runs roughly $17.6B in annual recurring revenue with more than 90% of revenue coming from subscriptions and a net revenue retention rate above 130%, according to a breakdown of Adobe’s subscription model. A retention rate over 100% means existing customers spend more every year, net of everyone who left. You do not get that number by trapping people. You get it because a designer’s entire body of work, muscle memory, file formats, and collaborators all live inside the tool. Leaving does not cost a migration fee; it costs a career’s worth of accumulated fluency.
Salesforce compounds the same effect through integration. Enterprises running ten or more Salesforce integrations show materially lower churn than those with a bare install, and a full CRM migration is commonly estimated to run 18 to 36 months once you count rebuilding automations, retraining teams, and re-plumbing every downstream system. The switching cost here is not friction Salesforce added. It is value the customer built on top of Salesforce and would have to build again.
This is the mechanism Nir Eyal describes in the investment phase of habit-forming products: every hour a user spends configuring, importing, and personalizing is stored value that raises the cost of leaving and, more importantly, raises the value of staying. The two move together. That is the tell of a durable switching cost. Friction only raises the cost of leaving; accumulated value raises both.
Retention economics explain why this is worth obsessing over. Bain & Company’s long-cited finding is that a 5% increase in retention can lift profits by 25% to 95%, because retained customers cost less to serve and buy more over time. But that curve only pays off when customers stay for value. Retention bought with friction inflates the top-line number while the underlying relationship rots, which is precisely the kind of hollow metric I have watched mislead a leadership team into thinking a product was healthier than it was.
What this looked like in the room
On a fractional COO engagement a few years back, I sat with a client running their operations on a billing platform the whole team openly hated. Clunky, slow, a support queue measured in days. Every quarter someone proposed switching, and every quarter the same reality killed it: three years of custom workflows, a dozen brittle integrations, and a finance team that knew the system’s quirks cold. The switching cost was real. We stayed.
But here is what the retention dashboard at that vendor could never see: we were not loyal, we were stuck, and we were actively shopping. The day a competitor showed up offering to rebuild our workflows for free and eat the migration, the conversation was over in a week. That vendor had a switching cost. It did not have a moat. It had a countdown, and it did not know the clock was running because we still showed up as “retained” every month.
The contrast that stuck with me came from a different tool the same client used, a data platform nobody complained about. No one proposed switching off it, not because leaving was hard, but because two years of dashboards, saved queries, and team habits made it more useful every quarter. Same category of switching cost on paper. Completely different relationship underneath.
The question that sorts the two
Before you count switching costs as part of your defensibility, run one test on each one: if leaving were instant and free tomorrow, would this customer still stay?
If the honest answer is yes, you have a durable switching cost. The customer stays for accumulated value, and no competitor buyout or regulatory mandate can hand that back to them, because it lives in their work, not your contract. Keep investing in it. Make the product measurably more valuable the longer it is used: more of their data organized, more of their workflow encoded, more of their skill and their team’s habits built on top of you.
If the honest answer is no, you do not have a moat. You have friction, and friction is on borrowed time. A funded competitor can price through it, a frustrated customer is already looking, and in a growing number of jurisdictions the law will pry the door open whether you like it or not.
This is why switching costs belong in the same strategic conversation as how you position against competitors who cannot afford to copy you and why pricing is itself a strategic act, not a footnote about retention mechanics. The strongest products do not make leaving painful. They make staying the obviously better deal, and they keep widening that gap every quarter the customer sticks around. Lock-in buys you time. Only accumulated value buys you a moat, and the difference is the whole game.