The Buyer’s Addendum, clause by clause.

Vendor agreements rarely define what “your data” includes, which means completeness gets defined later, by whoever performs the export. The five layers close the gaps where real migrations actually fail: a delivery of client demographics and invoices is not a medical record migration, and attachments and metadata are where exports quietly come up short.

The exclusion at the end is deliberate. The vendor keeps its software, templates, and configurations - the addendum claims your records, not their product. That boundary is part of why the definition is reasonable to sign.

The most common export failure is not refusal - it is a pile of PDFs. Technically your records, practically useless to the system you are moving to. This definition names that trick and rules it out.

The data dictionary requirement is the practical half: a structured export without documentation still forces your next vendor to reverse-engineer field meanings, billed by the hour. “Without the Vendor's assistance” is the test to keep in mind - if the outgoing vendor must be hired to interpret the export, it was not machine readable.

The numbers are starting positions: fifteen business days is enough time to run a real extract, and two free requests a year covers the normal case - one to evaluate a move, one to execute it - without letting exports become a harassment tool.

The last sentence is the one that matters most and gets noticed least. Some agreements start a deletion clock at termination. Without the suspension, a slow-walked export can run out that clock and the records are simply gone. With it, the vendor cannot delete data it has not yet delivered.

You should not need to open a support ticket to get a copy of your own records. A self-serve export removes the vendor's queue - and the vendor's mood - from the process entirely.

“Material term” is doing legal work here: if the export function disappears in a product update, that is not a roadmap decision you have to live with, it is a breach your attorney can act on.

This is acquisition-proofing. Export quality is rarely removed on day one - it erodes: a field dropped here, a format downgraded there, an export moved behind a support request. One sentence makes the capability you signed for the capability you keep.

The market default is drifting toward “we may de-identify, aggregate, and use your data to improve our services” - language broad enough to cover selling benchmarks and training AI models on your clinical records. This clause flips the default: the vendor may use your data to serve you, and everything beyond that requires consent you affirmatively give and can revoke.

The final sentence is the enforcement mechanism. Data-use grants rarely live in the agreement you read - they live in a data processing addendum or a privacy policy incorporated by reference and updated at will. This clause reaches those documents and supersedes them for Practice Data.

Note what it does not do: it does not prohibit the vendor from using your data with your consent. A practice that wants to participate in benchmarking or research can simply say yes - the point is that yes becomes a decision, not a default.

The asymmetry is intentional and it is what makes the clause signable. Reading your own data should not carry a recurring toll - it is your data. Write access genuinely costs the vendor engineering and support effort, so the addendum leaves it “as documented and priced” rather than demanding it free.

The deprecation notice protects the integrations your practice quietly depends on - labs, payments, communications tools. Twelve months converts “the API you built on disappears next quarter” into a planned transition.

This converts sales-call claims into contract terms. “We have an open API” and “we support standard formats” are cheap to say; a published page incorporated into the Agreement by reference is a representation the vendor is accountable for.

The second sentence is future-proofing for the industry. It names no specific standard - it simply says that if an open veterinary data exchange specification with a public conformance test exists, conformance claims must be verifiable by that test. No vendor can fail this clause by the ecosystem not existing yet; every vendor is held to it once it does.

Migrations do not usually fail because the data could not be moved. They fail because nobody defined what “done” meant, the go-live date arrived, and the practice went live on a partial conversion discovered exam by exam over the following months.

Two test migrations make the second one meaningful - the first finds the problems, the second proves they were fixed. Record counts by layer plus a spot-check protocol make acceptance objective, and the written sign-off gate means the go-live date moves when the data is not ready, instead of the data problem moving into production.

The defect-pays rule is symmetric on purpose: if the outgoing vendor delivers a broken extract, they pay; if the incoming vendor botches the conversion, they pay. Neither can bill you for redoing their own mistake.

An outgoing vendor has no commercial reason to be helpful, and the pricing of departure assistance tends to reflect that. This clause sets the terms while the relationship is still good: a response deadline, a rate card fixed in Exhibit A on the day you sign, and no minimum-purchase packages standing between you and a two-hour question.

In the months after a switch, someone at the front desk will need to check the old system - a record that converted oddly, an attachment that did not come across, a payment history question. Ninety days of read-only access is the practical safety net that keeps those moments from becoming emergencies.

The second sentence is the actual protection. Post-termination access priced at the moment of departure is an exit toll, quoted when your leverage is lowest. Fixing the fee at execution - years before anyone is leaving - is the entire point.

The auto-renewal trap has a standard shape: the decline window closes before you ever see the renewal price, and by the time the invoice arrives you are committed to another multi-year term. Every sentence here dismantles a piece of that - notice must arrive early and state the price, the decline window cannot be narrowed away, and renewals drop to one-year terms after the initial commitment.

The last sentence removes the incentive to “forget”: late notice does not shrink your decision window, it extends it. A vendor gains nothing by sending the renewal letter late.

“Renews at then-current list rates” is an unlimited price increase wearing boilerplate. The cap replaces it with a formula: CPI plus two points in normal years, never more than seven percent even when inflation spikes. The lesser-of structure protects you in both directions.

This is also the clause vendors most often want to negotiate, and that is fine - Exhibit A explicitly allows a different formula. A vendor proposing CPI plus three has engaged with the mechanism. A vendor refusing any cap at all has told you what their renewal pricing strategy is.

Consolidation is the veterinary software industry's normal state, not an edge case. The first sentence means the deal you signed survives your vendor being acquired - at least through the current term - rather than resetting to the acquirer's preferred pricing.

The second sentence is the sunset plan. Products get discontinued after acquisitions; when that happens, twelve months of runway, a free final export, and rate-card migration help are the difference between an orderly transition and a forced march on the acquirer's schedule.

This is the load-bearing sentence of the entire document. An addendum that conflicts with the main agreement and loses is decoration - and vendor templates sometimes contain their own precedence language designed to win exactly this fight. It is the first thing the vendor's counsel will read, and the cover note asks your attorney to confirm the vendor's paper does not defeat it.

The rights this addendum adds are mostly rights you need after the relationship ends - export, outbound cooperation, read-only access, discontinuation protections. Without a survival clause, contractual obligations can end with the agreement, which would extinguish these exactly when you need them. The list is precisely the exit machinery.

Minimalism is the negotiation strategy. The addendum does not touch the vendor's license grant, warranty, liability, or venue provisions - the clauses their legal team cares most about defending. The less it rewrites, the harder it is to reject wholesale, and the more a refusal says about the specific terms the vendor wants to keep unwritten.

Every mechanism in the addendum that involves money points here, and the table is filled in while you still have leverage - before signature, alongside pricing. A right to migration assistance at “then-current rates” is not a right; it is an invitation to be quoted later, under duress.

The exhibit is also a fast read on the vendor. Reasonable numbers written down promptly is the good sign it looks like. A vendor who will not put numbers in the table is telling you those numbers are meant to be decided after you are locked in - which is information worth having before you sign.