What happens to your data when you leave your booking software

A disclosure first, because this piece is unusual: we make booking software, so this is a guide to leaving written by a place you might one day leave. It covers leaving us too, and it has to — a page about data portability that quietly exempted its own author would be worth nothing.

Here is the position most people are in. You bought booking software because you wanted online bookings, and the question of what happens when you stop ranked somewhere between twelfth and never. Then something changed — the bill grew with the team, a commission landed that stung, a feature you needed did not arrive, or the business simply outgrew the tool — and now it is the only question that matters. You suspect you are stuck, and what you want to know is how badly.

The answer depends far less on the software than on two documents you have probably never opened and one thing you could do this afternoon.

What you actually own

"My data" is a vague noun, and vendors and customers use it to mean quite different amounts. Portability is not one thing that is either present or absent; it is six things, and they come out of a booking system in roughly this order of difficulty.

The customer list with contact details. Names, numbers, email addresses, and whatever you have tagged them with. This is what most owners mean when they say they want their data, and it is the one thing nearly every vendor will hand over, usually as a CSV. It is also the least of what you built.

Appointment history. Who came, for what, when, with which staff member, at what price. This is what turns a list of strangers into a business: it tells you who your regulars are, what your average ticket is, which service pays the rent, and who has not been in for eight months. It exports far less reliably than the list, and it is often split across several reports that do not join back up cleanly.

Notes. The free-text field where the real knowledge lives — the formula, the allergy, the dog that bites, the client who always runs fifteen minutes late. Notes are frequently attached to appointments rather than to people, which means an export that gives you customers but not appointment history gives you no notes at all.

Payment records. What was charged, taken, refunded, still owed. You may need these for tax long after the subscription ends. They are also the one category where a second copy usually exists: your payment processor holds its own record and its own export, independent of the booking tool, which is worth remembering when a platform's payment report looks thin.

Forms and documents. Intake forms, consent forms, waivers, questionnaires, signed agreements. These are stored as files rather than as rows, which is exactly why they go missing from exports. A CSV can hold an answer; it cannot hold a signed PDF.

Photos. Before-and-afters, colour records, condition shots on a car, progress pictures. Almost never in the export. Photos are large binary files and exports are text, and the gap between those two facts has cost more businesses more history than any deliberate lock-in ever has.

Read that list downwards and you have a map of what leaving costs. The top exports nearly everywhere, the bottom nearly nowhere, and most arguments about being held hostage are really about the middle.

Three questions to answer before you need the answers

The mistake is asking these on the day you decide to go, when you have no leverage and possibly no access. Ask while you are a happy paying customer, when the answers cost nothing and the vendor has every reason to be helpful.

1. What is the notice period, and what is it attached to

Find the terms of service — not the pricing page, not the FAQ, the actual agreement — and search it for "term", "renew" and "notice". You are after two facts: how long the commitment runs, and how far ahead of its end you must say something.

The strictest terms our competitive research found belong to Workiz, a field-service platform. According to Workiz's own terms and conditions, section 18.1, the agreement runs for an initial term of twelve months and then renews automatically for successive twelve-month periods unless one party notifies the other in writing no less than thirty days before the current term expires. Section 18.4 covers the case where you are the one leaving:

Pro may terminate this Agreement for convenience or for any other reason, at any time, by sending a termination notice to cancel@workiz.com, in which case termination will take effect at the end of the then-current Term, and Pro will not be charged for the subsequent Term. For the avoidance of doubt, any pre-paid Fees will not be refunded and all amounts under the original Term and Agreement will remain due.
— Workiz terms and conditions, section 18.4, read 18 August 2026

The trap there is not the notice period. It is the interaction between the notice period and the renewal. Miss the thirty-day window by a week and you are not late by a week — you are committed for another twelve months, and the fees for that year remain due whether or not you open the software again. That is a decision with four figures attached, made by a date in a calendar you have never looked at.

Two things belong in Workiz's favour, because a clause quoted without its neighbours is a misquote. Section 18.2 provides that if Workiz terminates for convenience it refunds a pro-rated share of pre-paid unused fees, so the obligation is not one-directional. And the terms are published in full on their own domain for anyone to read before signing — several vendors in the same tier publish no master agreement at all, which is a worse answer to this question, not a better one.

Most vendors are not this strict, and plenty advertise month-to-month terms on the pricing page because it sells. The point is that you cannot tell which kind you are signing without looking, and the look takes four minutes.

2. What happens to the data on the day the account closes

This is the question almost nobody asks, and it has the widest range of answers in the category.

At one end sits a promise of nothing in particular. Workiz's terms state that on termination the customer immediately loses all access to and use of the service, that the customer is responsible for downloading or backing up their own content, and that Workiz may permanently delete that content from its databases and servers. No window is stated. Notice what that combination does: it puts the duty to export on you and ends your access at the same moment. Each half is reasonable alone. Together they mean the export has to have happened already.

At the other end sits Cliniko, a clinic platform, and it is the most instructive case here — because its two published documents do not agree, and the disagreement is the single most useful thing in this piece.

Its help centre article on cancelling says the account drops into a limited-access, read-only state that administrators can still log into:

Cliniko keeps your account data for at least one year after cancellation. Even though the account is cancelled, you can log in during this period to export your data as well as to reactivate the account.
— Cliniko help centre, "Cancel your Cliniko subscription", read 18 August 2026

Its terms of service say something else:

All of your content and data will be deleted 90 days after cancellation or termination of your account. This information will not be able to be recovered. We recommend you use the export functionality of the Service prior to cancellation to retain any information you may require.
— Cliniko terms and conditions, read 18 August 2026

Both documents are Cliniko's, both are current, and the retention periods they describe differ by nine months. We are not going to guess which governs, and neither should you — but the lesson generalises to every vendor in the category. The help centre describes the intention; the terms of service are the contract. When they disagree, the contract is what would be argued in a dispute, and the shorter number is what to plan against. In this instance both answers are more generous than "may permanently delete", and the help article actively tells you to run your exports, so this is a discrepancy to notice rather than an accusation to make.

IntakeQ, which sells practice management to clinics, has done something more unusual: it has priced an answer. According to its help centre article on cancelling a subscription, a cancelled account loses access on its renewal date and the data is kept for thirty days after that — and instead of cancelling you are offered a read-only tier, pro-rated at the end of the billing cycle to a fee of $10 per month.

3. Run the export now, while you are still a happy customer

This is the practical heart of it, and it takes an afternoon.

Do not ask support whether an export exists. Go and find it, run it, and open the file. Then check three things: whether the appointment history came with it or only the customer list, whether the notes are in there, and whether there is any route at all to the uploaded forms and photos. Whatever you find is what you will have on the day you leave, and it will not improve under pressure.

What you learn from doing this is often more informative than the terms. Cliniko publishes a plain help article showing where the export lives, opening with the sentence "Your data always belongs to you, and you can download any of it for your own record keeping." That is a vendor answering the question before it is asked.

Jobber publishes an equally clear article on exporting clients — CSV or vCard, filterable by tag, admin users only, arriving by email — and it contains one sentence worth finding in advance: according to that help article, client exporting is available on select plans. We are not going to name which, because the article does not, and a guess printed here would be exactly the sort of unsourced claim this site refuses. Check it against the plan you are actually on. If your export lives one tier above you, today is a much better day to discover that than the day you cancel.

And then there is the shortest answer we found anywhere. Booksy's business help centre has an article titled "How do I cancel my Booksy account?" Its complete body, read on 18 August 2026, is an instruction to contact the support team at an email address. No notice period, no retention window, no mention of an export. That may be perfectly fine in practice — support may answer within the hour and do the right thing — but it is not published, so it cannot be checked in advance, and a thing you cannot check in advance is a thing you are trusting rather than knowing.

An export is not a migration

Here honest advice has to disappoint you slightly.

A CSV of customers is a mailing list. It is not a business. What actually takes the time when you move is everything that was never in the export: your service menu with its real durations, the extras and add-ons, the price that differs between the senior stylist and the junior for the same cut, deposit rules that vary by service, staff rotas and holidays and the one Wednesday in three that somebody works, the intake form with its conditional questions, the reminder schedule, the cancellation policy, and the packages a customer has already paid for and half used.

None of that is in the file. And even where the receiving system supports every one of those concepts, no two systems model them identically, so the mapping is a series of small judgements a person has to make. We know the size of that job from the inside, because we have run a real migration off another product's database — not a CSV, the whole database — and the extraction was the easy part. What cost the time was reconciling money that had been adjusted at the counter, deciding what to do about two customer records sharing one email address, and preserving the appointment numbers a business had been quoting to its clients for years.

So the useful reframing is this: portability protects your history, not your setup. The customer list, the appointment record and the payment history are the irreplaceable half — nobody can reconstruct those, and losing them is permanent. The configuration is the recoverable half: it costs a weekend, it is genuinely tedious, and it can be rebuilt from scratch by somebody who knows the business, usually better the second time because you will not carry over the workarounds.

Anyone who tells you switching booking software is a one-click import is selling you something. Anyone who tells you it is impossible is selling you something too.

What a vendor may legitimately hold on to

Not every limit on portability is lock-in, and a guide that treated them all as villainy would mislead you.

Clinical and health records sit under professional and legal retention rules that outlive your subscription and sometimes outlive your practice. A platform that keeps them for a defined period, and will not erase them the moment a card is declined, is doing right by the patient and usually by you. It is also why a paid read-only tier is a sensible product rather than a toll.

Payment and tax records carry retention duties of their own, on the vendor as much as on you. A platform that will not purge its ledger on request is usually obeying the same rules your accountant does.

Other people's data is the one that catches operators out, and it is sharpest with marketplaces. A consumer who downloaded an app, made an account and then booked with you has a relationship with that platform as well as with you. Their account, saved cards, review history and bookings across other businesses are not yours to export, and no amount of arguing about "my client list" changes it. What you can reasonably expect is the contact details of the people who became your customers, and the record of what they bought from you.

The marketplace question, honestly

Which leads to the concession a page like this must make or forfeit its credibility.

If a marketplace brought you clients, it brought you clients. That is not a rhetorical move, it is the actual product, and it works. A new shop in a city where nobody knows the name gets bookings from a consumer app that it would not otherwise have had, and an empty Tuesday costs more than a commission does. Their strength is real discovery infrastructure with a consumer audience already standing on it, and no standalone booking tool — ours emphatically included — matches it, because we have no consumer app and will not bring you a single new client.

So leaving a marketplace costs something portability does not refund. You can take the contact details of the people who became your customers. You cannot take your position in the app, the reviews attached to your listing, or the flow of strangers who were going to find you next month. Pick them, or stay with them, if new-client flow is still the binding constraint on the business — a fee for an introduction is a price you can reason about, and an empty chair is dearer than any commission. Revisit it when your book is full of regulars and the commission has stopped buying you anything you were not going to get anyway. We went through those fees vendor by vendor in our guide to the fees around Booksy and its alternatives, each figure read from the vendor's own pricing page.

What a good answer looks like

Here is the standard to hold any vendor to, this one included. None of it requires generosity — only specificity.

Two rules of thumb go with it. Prefer a number in the terms of service to a reassurance in an email, because only one of them survives a change of support staff. And if the export is gated behind a plan tier, read that as a fact about the vendor's posture rather than a pricing detail — a company that charges for the door does not think of the room as yours.

What we do

Late in the piece deliberately, and stated flat.

Reading and exporting your own records is not something a plan can withhold here, and that is a rule written into the code rather than a promise made in marketing. The comment above our capability catalogue reads: gate what a plan lets a business do; never gate seeing or exporting what is already theirs. In practice, appointment lists, customer lists, the CSV exports and the per-customer data export stay open on every plan, on the free plan, and after a subscription lapses. A business whose plan has ended can still see its data, take it away, and come back by paying. We took that decision because a booking system holding a customer list hostage over an unpaid invoice is not a thing we were willing to build, and because the legal right of access does not pause for a billing failure.

Concretely: appointments, customers and payments each export as CSV; there is a per-customer document containing everything held about one person, available both to you and to that person from their own account; and the whole record is readable programmatically through our API with a key you issue yourself. Cancelling is a form and a typed confirmation, you keep access to the date you have paid through, and there is no phone call and no retention queue. That is set out on our refunds and cancellation page, which you should read rather than take from here.

Now run our own checklist against us, because the standard above applies here too and we come off it with marks against our name.

The worst of them: our terms page is a plain-words summary, and it says in its own second sentence that the full terms are being prepared with counsel. So the rule of thumb one section above — prefer a number in the terms of service to a reassurance — currently indicts us. What we offer today is the reassurance. Related, and by our own table a poor answer: neither that page nor our privacy summary states how long your data survives after cancellation. It should, with a number, and it does not yet.

The rest are smaller. We publish no read-only archive tier, so long-term retention after you stop trading means your files on your disk rather than a login you keep. There is no single button that produces one archive of everything — it is several exports, run one at a time. Uploaded files and photos come out through their own routes rather than riding along in the CSVs. And we are a young product, so our exports have been used in anger far less often than an established vendor's, which is a real difference in evidence even where the intent is better.

If you are choosing between tools rather than escaping one, the wider method is in our eight questions for choosing booking software — question seven is this one, asked before you sign rather than after.