/ 9 min read
TL;DR
When you stop paying a SaaS vendor, your data gets locked rather than deleted, creating a gap between legal ownership and actual access that reveals the true cost of vendor lock-in during the departure process.
Your data doesn't get deleted the moment your card declines. It gets locked. There's a difference, and the difference is where most companies get hurt. You still legally own your records. You just can't reach them without paying. That gap between ownership and access is the whole story of SaaS lock-in, and it only becomes real when you try to leave.
Here's what happens in practice, what you actually get back, and what never comes out no matter how nicely you ask.
Most vendors follow a similar sequence. First comes a dunning email. Then a banner in the app. Then a grace period where everything still works. After that, your account flips to read-only or gets suspended entirely.
Read-only is the friendly version. You can log in and look at your records. You can't create, edit, or run automations. Sometimes you can still export. Sometimes export is the first thing they turn off, because export is the feature that makes leaving easy.
Suspension is worse. Users get a login wall. Integrations start returning errors. Any system you built that calls the vendor's API stops working, and if that API feeds your invoicing or your customer portal, your problem just got bigger than a subscription.
Deletion comes last. Retention windows after cancellation vary a lot by vendor and by plan. Some hold your data for a month. Some hold it longer. Some hold backups even after they've purged your live account, which is a compliance detail, not a recovery option. The only way to know your number is to read your agreement. Do that before you need it.
Several platforms restrict bulk export to higher-tier plans or to admins on an active subscription. If you downgrade to save money before you export, you can accidentally lock yourself out of the one feature you need.
You'll get files. That part is real. What you won't get is your system.
A typical export produces one CSV file per object. Contacts in one file. Deals in another. Projects, tasks, time entries, invoices, each in their own. Open one and it looks fine. Open two and try to connect them, and the trouble starts.
The link between records lives in ID columns. Your task file has a project ID. Your project file has a client ID. Those IDs are the vendor's internal numbers, and they mean nothing anywhere else. Rebuilding the relationships is real work. It's not hard work for a developer, but it's not a drag-and-drop into a new tool either.
Formulas come out as values, not formulas. A calculated margin field exports as the number it happened to be that day. The logic behind it is gone. Same with rollups, conditional formatting, and anything that computed on the fly.
Custom fields usually export, but the field definitions don't. You get a column called "Priority Tier" full of text. You don't get the dropdown options, the validation rules, or the fact that only three roles could edit it.
Some things almost never survive an export. Plan around that.
The pattern is simple. Rows come out. Behavior stays in. Your data is the easy part. The years of accumulated logic your team relies on every day is the part you'd be rebuilding from memory.
Almost every SaaS contract says you own your data. Read it closely. It says you own the content, not that the vendor guarantees you a usable copy of it in a format anyone else can read. Ownership and portability are two different promises.
Big exports take longer than you'd think. Vendors limit how fast their API can be called, usually to a few hundred requests per minute. If you have half a million records plus attachments, that's days of running, not an afternoon. Add retries for timeouts, and it stretches further.
Then there's shape. A tool that stores addresses as one text blob exports one text blob. Your new system wants street, city, state, and postal code as separate fields. Somebody has to split those, and the messy ones need human eyes.
Dates are a classic trap. Some platforms export timestamps in UTC, some in the account's time zone, and some inconsistently across objects. If your billing depends on which day work was logged, that's not a cosmetic issue.
Duplicates surface too. Every long-running system carries them. Migration is when you finally see how many. Budget time for cleanup, because importing dirty data into a fresh system just moves the mess.
Four clauses matter more than the price. Find them.
The first is termination and data retention. How long after cancellation can you still log in? How long until deletion? Get a number, in writing.
The second is export rights. Does the agreement promise machine-readable export of all your content, including files? Or does it just say "reasonable assistance," which means whatever they decide it means?
The third is suspension for non-payment. Some agreements let a vendor cut access during a billing dispute. If you're in a fight over an invoice, you don't want your operations as the bargaining chip.
The fourth is price escalation. Look for caps on renewal increases. A tool at a comfortable rate today can double over a few renewals, and by then switching is expensive because you're locked in by habit, not by contract.
You have the most leverage before you sign. Ask for a data portability clause with a defined format and a defined window. Vendors who intend to keep you honestly will agree. The ones who fight it are telling you something.
Run a full export once a quarter and store it in your own cloud storage. Do it while everything works. You'll find out what your export actually contains at a calm moment instead of during a cancellation, and you'll have a real backup if the vendor has an outage or a billing mistake.
Subscription cost feels small monthly and looks different across five years. Do the math on paper.
Say you have 45 people on a platform at 90 dollars per seat per month. That's 4,050 a month, or 48,600 a year. Add a 7 percent annual increase and five years runs past 280,000. Add the paid connector you need for accounting, the premium support tier, and the consultant you hire every time you want a new workflow, and the real number climbs further.
At the end of those five years, you own nothing. Stop paying and you're back to CSV files and a rebuild.
A custom system flips that. You pay to build it once. You pay to host it, which for most mid-sized operations is a modest monthly cloud bill, not a per-seat toll. Adding your 46th employee costs nothing. The source code is yours. The database is yours, running in your own cloud account, and you can query it, back it up, or move it any time without asking permission.
The break-even depends on your seat count and how weird your process is. Small teams on a cheap tool rarely clear it. Once you're paying five figures a year for something that still doesn't fit how you work, and you're paying humans to bridge the gaps with spreadsheets, the arithmetic usually favors building.
Building isn't automatically the smart move. Some honest limits.
If a commodity tool does the job well, keep it. Nobody should build their own email, payroll, or accounting ledger. Those are solved, regulated, and cheap. Custom work earns its keep on the process that makes you money, the part no vendor understands.
Custom software still needs care. Servers need patching, dependencies need updating, and someone has to answer when a user hits a bug. Owning the code means owning the maintenance. If nobody inside your company will ever touch it and you don't want an ongoing relationship with whoever built it, you've traded one dependency for another.
Migration is genuinely hard, and I won't pretend otherwise. Pulling twelve years of attachments out of a legacy platform is a scripted job. Retraining a team on new screens takes patience. If your clients log into the vendor's portal directly and like it, replacing that experience is a real cost, not a footnote.
And custom builds fail when the customer can't describe their own process. If three managers each run the workflow differently, software won't settle that. Fix the process on paper first.
Start with an inventory. List every SaaS tool, what it costs per year, how many seats, and what breaks if it disappears tomorrow. Most operations leaders are surprised by the total.
Then pick your two most critical systems and run a full export. Open the files. Check whether attachments came through. Check whether the relationships between records survived. That test tells you exactly how exposed you are, and it costs you an afternoon.
Finally, look at the tools where you're paying the most and fitting the worst. Those are the candidates for a system you own outright. If you want a second set of eyes on the math, book a free strategy session and we'll walk through your stack with you.
Questions
Your data gets locked, not deleted. You still legally own your records, but you can't access them without paying. The vendor typically sends a dunning email, displays a banner in the app, and gives a grace period where everything works. After that, your account becomes read-only or gets suspended entirely, preventing you from creating, editing, or running automations.
You'll get CSV files organized by object type, but not your system. Each file contains data with ID columns that are the vendor's internal numbers and meaningless elsewhere. Formulas export as static values, not formulas. Custom field definitions, dropdown options, and validation rules don't export. Rebuilding relationships between records requires developer work.
Attachments and uploaded files (most exports give expiring URLs, not actual files), automations and workflows, dashboards and reports, permissions and role structures, comment threads and activity history, and audit logs. Essentially, rows of data export but the accumulated logic and behavior your team relies on stays trapped in the platform.
Vendors limit API call speeds to a few hundred requests per minute. Exporting half a million records plus attachments takes days, not hours. Add retries for timeouts and data cleanup for formatting issues, and the timeline stretches significantly. Time zone inconsistencies and duplicate records require additional processing.
Four clauses matter most: termination and data retention (how long after cancellation until deletion), export rights (whether the contract guarantees machine-readable export of all content including files or just vague "reasonable assistance"), suspension for non-payment (whether a vendor can cut access during billing disputes), and any restrictions on bulk export or plan-based limitations.
No. Several platforms restrict bulk export to higher-tier plans or to admins on active subscriptions. Downgrading to save money before export can accidentally lock you out of the export feature entirely, which is the one capability you need to leave.
Data shape mismatches occur when one tool stores addresses as a single text blob but your new system needs separate street, city, state, and postal code fields. Date formats vary. Some export UTC, others use account time zone, and some are inconsistent across objects. Duplicates accumulate in long-running systems and surface during migration, requiring cleanup before importing into a fresh system.
Read this next
Getting ten years of records, attachments, and history out of a SaaS platform is doable, but it's the part most teams underestimate. Here's how the work really goes and what breaks.
If this is your situation
Thirty minutes over video with the person who would run the build. You talk, we ask questions, and you leave with a plain read on whether custom software makes sense for you. If the answer is that your current tool is fine and badly configured, you will hear that.
Who wrote this
Founder, Black Badger Software Solutions
David runs Black Badger from Clearwater, Florida. He has spent years implementing the platforms this site writes about, which is why the writing here is specific about where they fit and where they stop fitting.
Drafted with AI assistance, then checked, edited, and approved by a person before publishing. The facts, opinions, and recommendations are ours.
Related
Custom software /
Custom software isn't always the answer. Here's the arithmetic, the warning signs, and the honest caveats that tell you whether to renew your platform or replace it.
Custom software /
Platform consultants push you toward the standard workflow because that's what their incentives reward. Here's how to tell which of your processes deserve protection and which ones should go.
An article can only go so far without knowing your systems. Book a strategy call, walk us through what is breaking, and you will get a straight read on whether custom software is the right answer.
