Most practice owners do not think about data portability until they need it. Then they discover their practice management software has opinions about which columns they are allowed to see.
We hit this recently with Covetrus Pulse. We needed an enriched patient list — IDs, names, breed, weight, age, microchip numbers, allergies, owner contact details, account balances, the basics. Nothing exotic. Just the data that already lives in our own records.
Pulse would not let us have it.
What Pulse will and will not give you
Pulse offers a Reports module. Inside that module there are pre-canned reports — patient lists, client lists, accounts receivable, vaccine due reports, and so on. Each one produces a CSV with a fixed set of columns chosen by Covetrus.
You cannot pick the columns. You cannot combine fields from two reports into one file. You cannot ask for "every field on every patient in a single dump." That option does not exist in the interface. If you want owner contact info and patient demographics and the last-visit date in one row, you get three separate CSVs that do not join cleanly because the join keys are not all exposed.
For a small practice this is annoying. For anyone trying to do real analysis — figuring out which clients have not been in for eighteen months and carry a balance, or auditing breed distribution against treatment outcomes, or building a recall list for a specific demographic — it is a wall. Your data is in there. You can see it on the screen. You just cannot get it out in a shape that is useful.
Why this matters more than it looks
There are three places this hurts, and none of them are about export tools.
You cannot analyze your business with someone else's column choices. Every practice runs differently. The metric that matters to a mobile equine practice is not the metric that matters to a small-animal clinic. When the vendor decides which fields appear in your export, the vendor decides which questions you are allowed to ask of your own data. Most practice owners do not realize this until they ask a question and discover the answer is not exportable.
Switching practice management software is artificially expensive. Every PMS vendor on the market quotes a "data migration" line item when you onboard, and that line item exists for a specific reason: getting raw data out of the old system is hard by design. The incoming vendor charges you for the labor of recovering data that should have been yours to hand over in an afternoon. This is not a coincidence. The cost of switching is the moat that protects the vendor from competition. The vendor is not motivated to make the moat smaller.
You cannot apply your own data to automation or AI. Every interesting workflow a modern practice could build — automated reminders driven by visit history, AI-assisted triage prep, demand forecasting for vaccine stock, churn detection on inactive clients — starts with the same precondition: clean access to your own records. If you cannot get the records, you cannot build the workflows. Your competitors who are on more open systems will. Eventually that gap shows up in your books.
The bigger picture
This is vendor lock-in, and it is industry-wide in veterinary practice management software. The pattern repeats across most major players. Your data lives inside their system, structured the way they chose, exposed only through the reports they decided to write. The software is rented. The data should not be.
We are not the first to point this out and we will not be the last. The argument is straightforward: the practice generated the data, the practice paid for the software that holds it, and the practice should be able to walk out with a complete copy at any time, in a format that downstream tools can actually use. Anything less is a tax on practice owners — paid in switching costs, in lost analytical capability, and in the negotiating leverage you do not have when you cannot credibly threaten to leave.
What we did
When the report module would not give us what we needed, we went and got it ourselves. The short version: we pointed a browser at our Pulse instance and walked through every patient and client profile, reading each page the way a human would, and writing the contents into spreadsheets.
The longer version is still simple. We used browser automation — a category of software designed for this exact kind of structured-read task — to log into our own Pulse account, navigate to each patient's edit page in turn, expand the sections that hold the data we cared about, and capture every visible field. The same approach worked for the client list. The whole thing is the digital equivalent of opening every chart by hand and typing the contents into Excel, except the computer does the typing and does not get tired.
The numbers. 1,878 patient records. 748 client records. Two CSVs, one with full patient demographics and clinical metadata, one with full client contact, billing, and account information. Total elapsed time, supervised by one operator: roughly 100 minutes.
The cost. Effectively zero. The tools involved — a headless browser, a few hundred lines of glue code — run on a laptop. There is no SaaS subscription required. There is no per-record fee. The economic asymmetry is striking: a single afternoon of work produced a dataset that the practice management vendor would have been happy to never let us assemble.
The caveat. This worked because the practice owner authorized it. We were extracting our own data from our own paid Pulse account, signed in with our own credentials, against pages we are entitled to view as the account holder. This is not a hack and it is not a workaround in any adversarial sense — it is a paying customer pulling out the data they already own from the system they already pay for. The fact that we had to do it this way is the point.
What this means for you
If you are evaluating a practice management system right now, ask the vendor — before you sign anything — exactly what a bulk export of your raw data looks like. Be specific. "Can I export every field on every patient, every client, and every visit in a single dump, in a format that joins cleanly across entity types, on demand?" The honest answer is short. If the vendor hedges, you have learned something important about how they think about your data.
If you are already on Pulse, or on any other system where the answer to that question is no, you are not stuck. The data is still recoverable. It just shouldn't have to be.
We built decaVet to be the kind of practice that owns its own infrastructure — its schedule, its records, its relationships, its data. The barn aisle is not a place where you outsource the things that matter. Neither is your patient list.