The Vendor Exit Playbook: leave any tool in 30 days without breaking the business

Nobody plans to leave a vendor. You plan to stay, right up until the renewal quote arrives 40% higher, or the product you bought gets acquired and gutted, or the outage runs into day three and your customers are the ones asking questions.

At that moment you discover the only number that ever mattered: not what the vendor costs, but what it costs to leave. And you discover it at the worst possible time, with no leverage, no plan, and a renewal date in eleven days.

Exit readiness is not pessimism about your vendors. It is the single cheapest source of negotiating leverage available to a small company, and it is the difference between a bad quarter and a bad year. This playbook is the whole system: how to score what you are locked into, the 90-minute drill that proves it, the 30-day extraction plan, the scripts, and the contract clauses to fight for before you sign anything.

Part 1 — The mental model: the lock-in ladder

Lock-in is not binary. Every vendor sits on a rung, and the rung — not the price — determines your negotiating position.

1. L0 — Commodity. Interchangeable, month-to-month, nothing of yours lives inside it. You could switch on a Tuesday afternoon. (A password manager. A domain registrar.)
2. L1 — Data resident. Your data lives there, but it exports cleanly and means something outside the tool. (Most accounting, most email marketing.)
3. L2 — Configured. Months of setup live in it — workflows, automations, permissions, custom fields. The data exports; the thinking does not. (CRMs. Project tools. Helpdesks.)
4. L3 — Integrated. Other systems depend on it. It is the identity provider, the billing source of truth, the thing eight other tools authenticate against. Removing it breaks things you have not thought about.
5. L4 — Embedded. Your product or your delivery runs on it. Customers would notice within a day. Switching is a project with a budget, not a task with an owner.

The mistake is assuming your spend and your rung correlate. They usually do not. The $29/month tool that quietly became your single sign-on is an L3. The $2,000/month tool everyone complains about is often an L1 you could have replaced two years ago.

Part 2 — The Exit Readiness Score

Pull your card statement and list every recurring vendor. Score each one 0–3 on five axes. Low is good. Fifteen minutes for the whole stack, and it will be the most useful fifteen minutes of your quarter.

• Data portability (0–3). 0 = full export, standard format, self-serve, today. 1 = export exists but is partial or ugly. 2 = export requires a support ticket or a paid tier. 3 = no real export, or the export is a PDF, which is not an export.
• Reconstruction cost (0–3). How many person-days to rebuild the configuration elsewhere? 0 = under one. 1 = up to three. 2 = up to two weeks. 3 = you genuinely do not know, which is a 3.
• Integration depth (0–3). Count the other systems that would break. 0 = none. 1 = one or two, non-urgent. 2 = several, including something customer-facing. 3 = it is the identity, billing, or data spine.
• Contract exposure (0–3). 0 = monthly, cancel anytime. 1 = annual with a clean end date you have in your calendar. 2 = annual with auto-renew and a notice window. 3 = multi-year, or you cannot find the contract, which is also a 3.
• Knowledge concentration (0–3). How many people can actually operate it? 0 = most of the team. 1 = two or three. 2 = one person. 3 = one person, and they are a contractor.

Sum each vendor. 0–4: free to move, use it as a negotiating benchmark. 5–9: switchable with a plan, fix the worst axis this quarter. 10–15: you are not a customer, you are a hostage. Those get the fire drill.

Part 3 — The fire drill (90 minutes, quarterly)

Scores are opinions. A drill is evidence. Take your highest-scoring vendor and run this, timeboxed, without telling the vendor:

1. Export everything (20 min). Not a sample — everything. Open the file. Does it contain the fields that matter, or just the ones easy to serialise? Attachments and file uploads are where exports quietly fail.
2. Read the contract (15 min). Find four things: the renewal date, the notice period, the price-increase clause, and the data-return clause. Put the renewal date minus the notice period minus 14 days in your calendar as a recurring event. That date is your real deadline.
3. Map the dependencies (20 min). Open the vendor’s integrations or connected-apps page. Every item on that list is a thing that breaks. Write them down. This list is always longer than you expect.
4. Price the alternative (20 min). Identify one credible replacement and get a real number — list price, migration support, whether they will cover your remaining term. Vendors do this routinely; ask.
5. Write the one-pager (15 min). What we would lose, what it would cost, how long it would take, who would own it. File it where the next person can find it.

You now hold something most companies never have: a written, dated, evidenced answer to “what happens if we leave.” Everything downstream — renewal negotiations included — gets easier.

Part 4 — The 30-day extraction

When you decide to actually go, the failure mode is always the same: running the old and new systems in parallel for months, paying twice, trusting neither. Timebox it.

Days 1–5 — Freeze and inventory. Stop new configuration in the old tool. Export everything again, dated. List every workflow it performs and mark each one: keep, drop, simplify. The average migration drags 30–40% dead weight across; this is your chance to leave it behind. Name one owner. Migrations without a single named owner do not finish.

Days 6–15 — Rebuild the keeps. Only the keep list, in the new system, in priority order. Do not attempt feature parity — attempt outcome parity. Wire the integrations from your dependency map, most customer-visible first.

Days 16–22 — Parallel run, one cycle only. Both systems live for exactly one full business cycle: one billing run, one week of support tickets, one sales week. Log every discrepancy. One cycle, not two. Two becomes six.

Days 23–27 — Cut over. Switch, announce internally, and put the old system in read-only. Keep it read-only for one billing period as insurance.

Days 28–30 — Close it out properly. Send written cancellation (email, not a chat widget, and keep the reply). Take a final export and store it somewhere the vendor cannot revoke. Revoke the vendor’s API keys and OAuth grants — almost nobody does this, and stale grants outlive the relationship. Confirm no card on file.

Part 5 — The scripts

The renewal call, 30 days before your real deadline. Not a threat — a fact:

“We’re reviewing this line ahead of renewal. We’ve tested our export, we’ve scoped the migration to [alternative], and we have a number from them. We’d rather stay — switching costs us time we’d rather spend elsewhere. What can you do on price and terms to make staying the obvious call?”

Then stop talking. The specificity does the work: you have named a real alternative, and you have demonstrated the export works, which is the one thing that tells a vendor you are not bluffing.

The data-return request, if the export is incomplete:

“Please confirm in writing what data we can export, in what format, and by when. Please also confirm the process for full data return and deletion at contract end, per [clause reference]. We need this documented for our records.”

Written requests referencing a clause move faster than complaints. If your contract has no data-return clause, that is the finding.

Part 6 — Clauses to win before you sign

Exit readiness is cheapest at purchase, when you still have leverage and they still want the logo. Ask for four things:

• A price-increase cap at renewal — a stated percentage, not “then-current list price.”
• A short notice window — 30 days, not 90, and no auto-renewal without written notice to you first.
• A data-return commitment — full export in a machine-readable format, available for 30 days post-termination.
• A termination-for-convenience option on multi-year deals, even one that costs you a penalty. A priced exit beats no exit.

You will not get all four. You will usually get two, and the two you get will be worth more than the discount you were arguing about.

Part 7 — The failure modes

• Confusing the export with the data. A CSV that omits attachments, history, or custom fields is a trophy, not a backup. Open the file.
• Bluffing without a drill. Vendors negotiate for a living and can hear an unbacked threat. Do the drill first; the confidence is audible.
• Migrating the mess. A switch is the only free chance you get to delete 30% of your workflows. Waste it and you have paid for a new home for the same problems.
• Forgetting the humans. The team who used the old tool for three years will be slower for two weeks. Budget for that, and say so out loud, or the migration gets blamed for a dip everyone should have expected.
• Leaving the doors open. Cancelled subscription, live API key. Revoke access on the way out, every time.