The scenario people do not plan for
You have used an app for three years. Your dog's full medical history is in it, or four years of trip records, or every measurement you have taken since you started.
Then one of these happens:
- An email announcing the service closes in 30 days
- An acquisition, followed by a very different product
- Nothing at all. It stops syncing. Support does not reply. One day it will not open.
The third is the most common and the only one with no warning. Given that the median app retains around 4 percent of its users after 30 days, most apps are commercial failures, and most commercial failures end quietly rather than with an announcement.
This is about what actually happens to the records in each case, and how to make sure it costs you nothing.
The three endings
A managed wind-down
The good version. A company that intends to behave well gives notice, usually 30 to 90 days, publishes an export tool, and states a date after which data is deleted.
Worth knowing: there is no general legal requirement to give notice before shutting down a consumer service. GDPR governs how data is processed and your rights over it, not whether a company must keep operating. The notice period is a choice.
What to do inside it: export immediately, on the day you get the email. Export tools during a wind-down are built quickly, run under load from everyone doing the same thing, and are often the first thing to break.
An acquisition
Your data usually moves to the buyer.
Nearly every privacy policy contains a clause resembling:
In the event of a merger, acquisition, or sale of assets, your information may be transferred as part of that transaction.
That clause is doing a lot of work. It means the policy you agreed to can be replaced by a different company's policy, with different sharing practices and a different business model. You are typically notified, and the practical options are accept or leave.
The pattern to watch: an acquisition by a company whose revenue comes from advertising, when the app you chose was funded some other way. The incentives around your data change even if the interface does not.
Abandonment
No announcement, because nobody is left to write one.
The sequence is recognisable. Updates stop. It breaks after an OS update and stays broken. Support goes unanswered. Sync fails intermittently, then permanently. The app remains installed and looks fine while quietly holding a stale copy of your data.
This is the common case, and it is the one with no export window, because by the time you are certain it is over the servers are already gone.
What "your data" actually means here
Three copies, with different fates.
| Where it lives | What a shutdown does |
|---|---|
| On your device | Survives until you delete the app. Often unreadable outside it. |
| On their servers | Inaccessible once the service stops. Deleted on their schedule, or not. |
| With third parties | Analytics and advertising partners already hold their copies, unaffected by the shutdown |
That third row is the one people never consider. If an app shared behavioural data with advertising partners, that data does not disappear because the app did. It sits in profiles held by companies you never chose. This is one of several reasons the sharing section of a privacy policy matters more than the rest, covered in how to read a privacy policy in five minutes.
Also worth being clear: deleting the app deletes nothing on the server. Removing it from your phone removes the local copy and your access. The account and its contents persist until you delete the account or submit a deletion request.
Exit cost, decided before you need it
The whole thing reduces to one property, best assessed on day one rather than during a wind-down.
Low exit cost. Export to CSV, JSON, PDF, or an open standard. You can leave in ten minutes and the file opens without the app.
Medium. An export exists but is partial, awkward, or in a format only the app reads.
High. No export at all. Your records are viewable only inside software you do not control.
High exit cost is a design choice, not a technical limitation. Writing a CSV is trivial. Its absence is usually deliberate, because data you cannot remove is data that keeps you subscribed.
The consequence: in two years you may be using something because leaving is expensive, not because it is good. That is captivity, and a shutdown converts it into loss.
What a good export looks like
"It has an export" is not enough on its own. Exports vary enormously in how much they actually preserve, and the difference only becomes visible when you need one.
Complete. Every record, not just the last 90 days or the current view. A surprising number of exports quietly cap themselves, and you find out by counting rows.
Structured. CSV, JSON, or a documented open format. A PDF is fine for handing to a professional and useless for moving into another tool, because you cannot import a PDF.
Self-describing. Column headers that mean something without the app to explain them. An export with columns labelled f1, f2, f3 is technically an export and practically a puzzle.
Includes the attachments. Photos, scanned documents, and PDFs are usually stored separately from the records that reference them. An export of the database with none of the files is half your data.
Runs without a subscription. Some services put export behind the paid tier, so cancelling locks your history inside an account you are no longer paying for. Check this before you subscribe, not after.
A quick test costs two minutes: run the export today, open the file, and count the rows against what the app displays. If they do not match, you have learned something important while it is still cheap to act on.
What to actually do
Before you commit
Check for an export function before you start building records. Not the marketing page, the actual settings screen. Search the app name plus the word export.
This matters most for anything a professional might need: pet medical history, health measurements, financial records, anything with dates you will be asked to recall precisely.
While you use it
Export on a schedule. Quarterly for anything you would be upset to lose. Put it in the calendar; it takes two minutes and it is the entire insurance policy.
Keep the export somewhere neutral. A CSV in your own cloud storage or a local folder. Not inside another app with the same problem.
Prefer formats that outlive software. CSV and PDF will open in twenty years. A proprietary backup file will open in whatever still supports it.
Watching for the warning signs
- No update in 12 months, especially after an OS release
- Support requests going unanswered for weeks
- Recent reviews reporting the same broken thing over months
- A free app with no visible revenue model and no stated plan
That last one deserves care. As covered in how free apps actually make money, an app with no funding model is not sinister, it is unresolved — and unresolved resolves eventually, into advertising, a subscription, or shutdown.
If it has already happened
- Open the app offline and screenshot anything critical. Local data often survives after servers stop.
- Look for a local backup or cache the app may have written to device storage.
- Submit a GDPR or CCPA data request if the company still legally exists. Obligations survive a product's closure even when the servers do not.
- Check whether the developer published an export tool or migration path, often only mentioned once on a support page.
Which records deserve the effort
Exporting everything quarterly is more discipline than most people will sustain, and it is not necessary. Sort what you hold into three tiers and treat them differently.
Tier one: records someone else will ask you for
Pet medical history, health measurements, financial records, anything a vet, doctor, accountant, or insurer might need. These have two properties that make them the priority: they are irreplaceable because they are a record of events that already happened, and they are needed under pressure, usually at short notice in a consulting room or an office.
Export these on a schedule and keep a copy in a format that opens without the app. This is the tier where a shutdown turns from annoying into genuinely costly.
Tier two: records with sentimental or reference value
Trip histories, photo collections, long-running logs you like having. Losing them would upset you and would not change any decision.
Export yearly. Do not agonise about it.
Tier three: operational state
Current checklists, active reminders, this week's plans. It looks important because you are using it right now, and it is the least worth protecting, because it expires on its own. Next month's version bears no relation to this month's.
Do not export this at all.
Most people get the tiers backwards. They worry about the live data they can see and never think about the four-year history quietly accumulating in the background, which is the only part that cannot be recreated.
Where we stand
We build apps, so this is a disclosure and includes the part that does not flatter us.
PackPilot and PawDex are young, and by the criteria above that is a legitimate risk you should weigh. Both are free with no advertising, which by our own framework in how to choose an app you will still use in a year puts them in the unresolved-funding category.
On exit cost specifically: PDF export of PawDex logs for vet visits is on the roadmap and is not shipped. Until it is, exit cost for PawDex is higher than we want it to be. That is the weakest point in either product and it is the one most relevant to this article, so it would be dishonest to write 1,500 words on data portability and leave it out.
Judge that against what ships, not against this paragraph.
The short version
- Most apps end quietly, without notice, because most apps fail.
- Deleting the app deletes nothing on their servers.
- An acquisition transfers your data under a clause already in the policy you accepted.
- Third parties keep their copies regardless.
- Check for an export before you build records, not when you need one.
- Export anything important quarterly, in a format that opens without the app.
A shutdown should be an inconvenience. It only becomes a loss when the data existed in exactly one place that somebody else controlled.
Related Reading
- How to Read a Privacy Policy in Five Minutes covers the sharing and retention clauses that decide this.
- How to Choose an App You Will Still Use in a Year covers exit cost as a selection criterion.
- How Free Apps Actually Make Money explains which funding models tend to end quietly.
- Subscription Creep: How to Audit What You Are Paying For covers the recurring charges that outlive their usefulness.
- How to Tell If an App Is Abandoned covers how to see it coming.
Frequently Asked Questions
It depends entirely on whether the company runs a wind-down. A managed shutdown gives notice and an export window. An abandonment gives neither: servers eventually stop responding and anything held only on them is gone.
A responsible wind-down gives 30 to 90 days. There is no general legal minimum, so the notice period is a choice rather than a requirement, and abandoned apps frequently give none at all.
It usually transfers to the acquirer. Most privacy policies contain a clause allowing data to be transferred as a business asset in a merger or acquisition, which means the new owner's practices apply, not the ones you originally agreed to.
No. Removing an app deletes the local copy on your device. Anything synced to the company's servers remains until you delete the account or submit a deletion request.
The right under GDPR to receive your personal data in a structured, commonly used, machine-readable format and to have it transmitted elsewhere. It only helps if you exercise it while the company is still operating.
Export them on a schedule rather than waiting for a reason. If the data would matter to a professional such as a vet, doctor, or accountant, keep a copy outside the app in a format you can open without it.
Free apps with no visible revenue model, apps from single-person developers with no update in a year, and apps whose parent company treats them as a side project rather than a business.