Switching Cost
Switching costs are everything a customer has to pay — in time, money, risk, and political capital — to replace one vendor with another. They are not the sticker price of the new tool. They’re the full bill for getting out of the old one and getting productive on the new one, including the parts that don’t show up on any invoice. Most operators discover switching costs the hard way, usually somewhere between the third migration deadline slip and the moment someone senior asks why the project is six months late. They are the single most underrated variable in B2B software economics, and they’re the reason most “cheaper and better” alternatives quietly lose to incumbents that are neither.
Why Switching Costs Are the Real SaaS Moat
Features get copied. UX gets copied faster. What does not get copied is the seven years a customer has spent configuring, importing, integrating, and training staff on the incumbent product. That accumulated investment is the moat. Every day a customer stays, the moat gets a little deeper, because more data flows in, more integrations get built, and more people learn the system’s quirks. When a cheaper alternative shows up, the buyer isn’t comparing two products. They’re comparing “keep paying” against “pay the new vendor plus eat the switching cost plus absorb the risk of something going wrong.” The math rarely favors switching unless the incumbent is actively failing.
This is why the “AI makes cloning software trivial” argument collapses on contact with real enterprise accounts. Cloning the software is the easy part. Cloning the embedded position is not a software problem.
Components of Switching Costs
Switching cost is almost always a stack of smaller costs that add up to something larger than any single line item. Data is the most obvious layer — exports that don’t round-trip cleanly, historical records in proprietary formats, fields that don’t map one-to-one, and the long tail of edge cases that only surface during migration testing. Integrations are usually the biggest hidden cost — every other system wired into the incumbent (BI stack, identity provider, ticketing, CRM, billing) has to be rewired, tested, and monitored, and each rewire is its own mini-project. Workflow embedding is the layer nobody documents — the muscle memory, the custom processes built around the tool’s specific behavior, the “we always do it this way because of how this field works.” Trained staff represents real hours of retraining, with real productivity loss during the transition. Audit trail and compliance history matters in regulated environments where the incumbent holds years of evidence that would be painful to reconstruct. Political capital is the quietest but often decisive factor — the person who championed the incumbent vendor has reputation on the line, and replacing it feels like admitting the original decision was wrong. That’s a human cost, not a technical one, and it reliably kills migration projects that make financial sense on paper.
Why “AI Makes Software Cheap to Build” Doesn’t Dissolve Switching Costs
The bullish case for AI-native challengers assumes a greenfield customer. That customer does not exist in most of the enterprise market. Real customers already have data, integrations, habits, dashboards, and dependencies sitting on top of the incumbent product. A challenger showing up with a cheaper or technically better clone faces a buyer who has to weigh the theoretical gain against a very real migration bill. In most cases, the incumbent only needs to be good enough, not best-in-class. Good enough plus embedded wins against better plus greenfield, because only one of those two ships has a compounding cost of leaving.
This is exactly the confusion at the center of the “SaaS is dead” narrative — code got cheaper to generate, but code was maybe 15% of what kept customers from leaving. The other 85% is switching cost in one form or another, and AI doesn’t touch it.
Strategic Implications
Incumbents with durable positions defend by deepening integration. More APIs, more partner ecosystems, more data surface, more workflow hooks, more places where leaving means breaking something. The goal is to make the cost of exit rise faster than the value gap with any challenger. Pricing is one lever in this — outcome-based pricing can further raise switching cost by tying the vendor’s revenue to measurable results the customer would have to re-baseline from scratch under a replacement.
Challengers win by refusing to fight on the incumbent’s ground. They target segments where switching cost is still low: new teams, greenfield deployments, adjacent use cases the incumbent doesn’t cover well, or emerging buyer personas who never adopted the incumbent in the first place. Once a challenger establishes a position in that beachhead, the same dynamic starts working in its favor — every month it holds the account, its own switching cost rises, and the moat starts to fill.
The operators who miss this tend to be the ones who overweight product quality in their strategic planning. Product quality matters. It is not, in B2B software, the dominant variable. The dominant variable is what it costs the customer to leave, and most of that cost is not in your product at all — it is in the rest of the customer’s business.
For the broader argument about where the AI pressure on SaaS actually lands and what it doesn’t touch, see Everyone Says SaaS Is Dead. Here’s What They’re Actually Observing.