What does it actually cost to leave?
Most "TCO calculators" make you guess a migration cost. This one doesn't. Pick the vendor you're leaving and it derives the one-time switching cost from that vendor's own lock-in profile, because the friction of leaving Oracle (proprietary SQL surface), Veeam (a backup-retention tail), and Okta (credential captivity) lives in completely different places. 98 vendors, each modeled on its own parameters.
| Switching work | What drives it | ~Weeks | Cost |
|---|---|---|---|
| Data extraction & validation | Data portability (D) | 7 | $28,000 |
| Re-platforming proprietary surface | Proprietary surface (P) | 8.8 | $35,200 |
| Team retraining & runbooks | Ecosystem & skills (E) | 4.2 | $16,800 |
| Cutover & parallel validation | Exit-path maturity (X) | 2.2 | $8,800 |
| Labor subtotal 22.2 engineer-weeks @ $4,000/wk | 22.2 | $88,800 | |
| Parallel-run overlap you keep paying Azure OpenAI while the project runs | $20,492 | ||
| One-time switching cost | $109,292 | ||
Illustrative estimate, not a vendor quote. The one-time cost is modeled from this vendor's Lock-in Index profile and your estate size; real effort depends on your architecture, freeze windows, and team. Every figure above is editable. Confirm pricing with the vendor or an authorized partner.
Why every vendor calculates differently
A flat "migration is 20% of annual spend" rule is worse than useless: it tells a team leaving PostgreSQL and a team leaving Oracle the same thing. This model instead reads the five Lock-in Index dimensions for the incumbent and maps each one to a distinct line of switching work:
- Data portability (D) drives data extraction & validation: can you get your data, backups, and identities out in a format something else can read?
- Proprietary surface (P) drives re-platforming: how much config and code targets vendor-only languages and features that must be re-expressed.
- Ecosystem & skills (E) drives retraining & runbooks: how much operational muscle memory assumes this product.
- Exit-path maturity (X) drives cutover & parallel validation: whether a mature tool exists or you're building the migration by hand.
- Licensing power (L) shapes the stay-put risk: the vendor's ability to reprice you at renewal, the cost of not moving.
Effort scales sub-linearly with estate size (doubling the estate does not double the re-platforming work), and a parallel-run overlap adds the cost of paying the incumbent while the project runs. The result is an honest shape, not a round number: two vendors at the same price can land quarters apart on effort.
How to read the result
Payback period is the one number a CFO will anchor on: how many months of the annual saving it takes to earn back the one-time switching cost. Under a renewal horizon it is often shorter than people fear, because the recurring saving compounds while the switching cost is paid once. Net over the horizon is savings across your chosen number of years minus that one-time cost, the go/no-go figure.
This is a planning skeleton, not a quote. When you're ready for binding numbers, the 3-year TCO calculator models support tiers and generates a decision report, the renewal planner tests whether the timeline fits before your contract forces the question, and each category's exit guide covers the tooling that actually does the move. Watch the Price Hike Tracker for the licensing moves that change these numbers.