Methodology & transparency

Where the data comes from

VetPulse doesn't replace your practice information management system (PIMS) — it reads from it. Once a clinic connects its PIMS or uploads a CSV export, VetPulse treats that data as the source of truth and derives every metric, alert, and briefing from it. VetPulse is not itself the system of record for scheduling, medical records, or billing; your PIMS remains that.

How metrics are calculated

Every KPI VetPulse surfaces — average transaction value, utilization rate, no-show rate, revenue per doctor hour, care compliance rate, and the rest — is calculated using a fixed, documented definition rather than an ad-hoc one. Those definitions are the same ones written up in the practice KPI glossary: if a metric's calculation ever changes, the glossary entry is the place that reflects it, so a number on a dashboard and its definition never drift apart.

Benchmarks referenced in VetPulse's blog contentare general industry ranges for context, not per-clinic guarantees — they describe what "normal" tends to look like across independent practices, not a target any specific clinic's numbers are being graded against.

How findings and alerts are surfaced

VetPulse runs a set of detection rules against a clinic's connected data on an ongoing basis, looking for the specific patterns each rule is built to catch — a lapsed client overdue for care, an unbilled procedure, a developing care gap, or inventory nearing expiry. When a rule matches, it becomes a finding: a scored, ranked item that shows up in the live alert feed and the weekly briefing, along with the underlying data that triggered it.

The weekly briefing bundles the week's most relevant findings into a plain-English summary delivered on a fixed cadence, rather than requiring an owner to pull and interpret a raw PIMS report themselves.

How to interpret an insight

Every finding VetPulse shows is meant to be checked against the underlying data it cites, not taken at face value — the goal is to point an owner at something worth a closer look, not to hand down an unreviewable verdict. A finding reflects a pattern in the connected data as of when it was generated; it is not a prediction, a diagnosis, or a substitute for professional judgment about a specific patient, client, or financial decision.

New in this release

Three new findings joined the detection rules described above.

New Client Decline compares new-client acquisition against the prior comparable period and flags a meaningful drop — a leading indicator, distinct from a lapsed client, which has already stopped visiting rather than never started. See New Client Decline Alerts for how this is surfaced.

Revenue Concentration looks at what share of total household lifetime spend sits with the clinic's top-spending clients — a structural retention risk that exists independently of any one client showing signs of lapsing. See Revenue Concentration Software for how this is surfaced.

Vet No-Show Disparity breaks the clinic-wide no-show rate down by provider, so a pattern specific to one vet's schedule doesn't stay hidden inside a single clinic-wide average. See Vet No-Show Disparity Alerts for how this is surfaced.

Two existing findings, Average Transaction Value Gap and Inventory Overstock, also moved onto the same real-time detection pipeline as every other finding in this release — previously they ran on a separate weekly schedule with their own, independently-implemented thresholds. Their underlying calculations are unchanged; only where they run changed.

Wherever two findings report a number that should logically agree — for example, a provider's no-show rate compared against the same clinic-wide average Missed Revenue reports, or a household's lifetime spend appearing in both the Revenue Concentration and Lapsed Clients findings — that figure is calculated exactly once and shared, not computed twice by two independent rules. That's a design constraint on how VetPulse's findings are built, verified automatically every time the detection logic changes, not just a claim made here.

New in this release (v23)

Peer Comparison Deviation compares a clinic's share of revenue for a single service category against the median for a cohort of similar-sized practices in the same region, rather than against the clinic's own history the way a simple revenue trend does. A comparison is only ever shown once its cohort clears two independent gates — at least 20 contributing practices, and no single practice contributing more than 40% of that cohort's revenue for the account/period in question — and every figure shown is an aggregate statistic (median, 25th/75th percentile), never another practice's individual numbers. See Peer Comparison Benchmarking for how this is surfaced.

What this page is not

This page describes VetPulse's own methodology, not an independent study, dataset, or industry benchmark report. Ranges referenced elsewhere on this site come from general veterinary business context, not from a formal VetPulse research dataset.