Glossary

The TechStack vocabulary.

Plain-English definitions for every term we use. If you've ever wondered what Drift Score, Schedule Value, or Empirical Bayes means in the context of running a practice — start here.

Book Value: The present value of the future revenue embedded in a practice's active client relationships.; Book Value is calculated as the sum, across every active client, of expected annual revenue × retention probability, discounted to today's dollars over a 3-year horizon. Unlike monthly revenue (which is backward-looking), Book Value is forward-looking — the number a private equity firm uses to price a roll-up acquisition or a bank uses to underwrite a line of credit. It moves daily as clients drift into or out of healthy cadence.
Drift Score: A 0–100 score representing how far a specific client has departed from their personal booking rhythm.; Drift Score is computed using empirical Bayes — blending the client's personal median inter-visit gap with a peer prior calibrated to their industry segment. Each score is paired with plain-English drivers (e.g., '9 weeks overdue on a 5-week personal rhythm; 2.5× typical variation'). Bands: 0–20 on cadence, 21–40 stretched, 41–60 drifting, 61–80 high drift, 81–100 likely gone.
Drift Radar: Per-client churn prediction system that scores each client against their own historical booking cadence.; Drift Radar replaces the generic dormancy threshold (e.g., 'flag at 90 days') used in most retention tools. Instead, it learns each client's personal rhythm and surfaces drift the moment that rhythm starts to crack — typically 6–10 weeks earlier than threshold-based systems. Calibrated with industry-specific peer priors for hair color, medspa tox, dental hygiene, monthly massage, and more.
Voice Training: A system that learns a practice owner's writing voice from past messages and generates outreach drafts in their tone.; Voice Training extracts features (greeting style, sign-off, average length, formality, warmth, emoji rate, distinctive phrases) from 15–30 messages an owner has already sent, then renders every reactivation, recall, or outreach draft in their voice — across three tone variants (your tone, a little warmer, a little more direct). Runs as a deterministic template engine; no LLM required.
Schedule Value: The dollar value of a service business's next 14 days of bookings — booked plus expected fill of open hours.; Schedule Value is the sister metric to Book Value, applied to time instead of relationships. For each day in the horizon, it sums (a) confirmed bookings × average revenue and (b) open slots × historical fill rate × average revenue, then flags slots running below 35% fill rate as 'recoverable' opportunities for targeted flash offers. Available in Revenue IQ.
Personal Cadence: An individual client's median inter-visit gap, with associated standard deviation, derived from their visit history.; Personal cadence is the foundation of per-client drift detection. A client who books every 28 days with a 6-day standard deviation has a fundamentally different cadence than one who books every 84 days with a 14-day std. Generic dormancy thresholds ignore this distinction; personal cadence respects it and surfaces drift relative to each client's own pattern.
Empirical Bayes: Statistical technique that blends a client's personal history with a population-level prior, weighted by how much data is available.; When a client only has 1–2 visits, you can't reliably compute their personal cadence — you'd be making a strong claim from thin data. Empirical Bayes handles this by blending the client's median gap with an industry-segment peer prior, weighted by the number of observed visits. With more visits, personal history dominates; with fewer, the peer prior dominates. The result is a stable, defensible estimate across the entire client base.
Peer Prior: An industry-segment baseline for typical visit cadence, used as a fallback when individual client history is thin.; TechStack ships peer priors for hair color (42-day median, 14-day std), medspa tox (105/21), dental hygiene (180/30), monthly massage (28/10), HVAC tune-up cadence (180/45), quarterly pest control (90/10), and more. The right peer prior depends on the segment label assigned to each client.
Dead Time: Hours on a practice's schedule that historically run below 35% fill rate.; Dead time is the inverse of demand: the time slots clients reliably don't book. Most practices have predictable dead-time pockets (Tuesday 10 AM, Thursday 1 PM, Saturday 4 PM) that haven't filled in months but never get specifically targeted. Revenue IQ ranks dead slots by recoverable revenue and drafts flash offers sized to the deadness of each slot.
Flash Offer: A time-bound discount sent to a specific client for a specific empty time slot, with the discount sized to the slot's historical dead-ness.; Flash offers are not blanket discounts. The deader the slot, the bigger the discount; peak-time slots never get discounted. Revenue IQ generates flash offers automatically — picking the client most likely to take the slot, drafting the message in the owner's voice, and applying a discount within the owner-set guardrails.
Demand Heatmap: A 7-day × hours-of-day grid showing booking density across a practice's typical week.; Visualizes which day-and-hour combinations consistently fill up (peak) and which run light (dead). Revenue IQ overlays a 35%-fill threshold to flag dead slots automatically. The visualization is owner-facing — most practices are surprised by where their dead time actually sits.
Recall: The dental industry term for a patient's scheduled return visit, typically a 6-month hygiene cleaning.; Every dental PMS (Dentrix, Open Dental, Eaglesoft, Curve, Denticon) has built-in recall flagging. The bottleneck isn't the flag — it's the operational follow-through. Practices running 85%+ recall compliance use a multi-channel cadence (postcard + email + SMS + phone) with personalized messaging per dormancy band, not generic 'time for your cleaning' postcards.
Attribution: Tying a recovered booking back to the specific outreach message that produced it.; Real attribution requires (1) per-client tracking tokens embedded in outreach links, (2) booking-side capture matching click → appointment by email + time window, and (3) revenue-side realization when the appointment completes. Most retention tools report 'sends' and 'clicks' but stop short of tying back to actual bookings or revenue. TechStack's attribution layer closes the loop end-to-end.
Lifetime Value (LTV): The total revenue a single client is expected to generate over the full lifetime of the relationship.; For service businesses: average ticket × visits per year × expected years of relationship. The 'expected years' factor is what most owners get wrong — they use industry averages instead of their own retention data. TechStack computes LTV per client using actual cadence and personalized retention probability, which produces more honest numbers than the typical generic formula.
Retention Probability: The probability that a specific client will still be active with a practice one year from today.; Derived from the client's drift score: drift 0 → ~97% retention, drift 50 → ~70%, drift 80 → ~30%, drift 100 → ~5%. The curve is sigmoid-shaped around drift 50. Retention probability compounds across years, which is why a single year of drift typically destroys 30–50% of a client's lifetime value.
BAA (Business Associate Agreement): A HIPAA-required contract between a healthcare practice and any vendor that handles protected health information (PHI).; Before patient data can move between a dental or medical practice and a software vendor, both parties must sign a BAA establishing each side's compliance obligations. TechStack executes a BAA after a healthcare-readiness review and before any PHI moves for dental, medspa, and medical practice customers; PHI is encrypted at rest and in transit, access is audit-logged, and outreach messages never reference specific procedure codes or treatment details.
PMS (Practice Management System): The core software a dental, medical, or veterinary practice uses to manage patient records, scheduling, and billing.; Major dental PMS platforms include Dentrix, Open Dental, Eaglesoft, Curve Dental, Denticon, Carestream, and PracticeWorks. Most retention tools (including TechStack) sit on top of the PMS via CSV export or direct integration — they don't replace it.
Evidence Grade: One of five labels — Confirmed, Attributed, Inferred, Estimated, Unavailable — describing how strongly a recovered visit can be tied to Retention IQ outreach.; Every attributed return in Retention IQ carries an evidence grade assigned by the Four-Gate Evaluator. Confirmed means all four gates passed; Attributed, Inferred, and Estimated each name a weaker, specific kind of evidence; Unavailable means no qualifying send exists. Only Confirmed returns can become billable line items, which is what makes pay-per-verified-return pricing auditable.
Four-Gate Evaluator: The rule set that grades a recovered visit by testing four gates: dormant, delivered, timed, and converted.; The evaluator asks four questions of every candidate return: was the client dormant when outreach began, was the message delivered, did the return fall inside the attribution window, and was the visit completed and paid. A return that passes all four is Confirmed. The evaluator ships with known-pass and known-fail self-tests so a change to grading logic is caught before it can change an invoice.
Enrollment Snapshot: An immutable record of a client's state at the moment they were enrolled in a campaign.; When a client is enrolled in outreach, Retention IQ captures their cadence, segment, identity match, consent status, suppression state, and qualification or exclusion reasons as a versioned snapshot that cannot be edited afterward. Evidence grades are computed against the snapshot, so the proof behind a billed return never drifts as the live client record changes.
Contact Exception: A permanent, channel-specific do-not-contact rule (contact, SMS, email, or call) enforced before any outreach, reply, booking action, or call.; Contact exceptions are practice-level and permanent until a person removes them, unlike per-campaign suppression lists. They are checked in front of every channel the platform touches and apply in addition to consent flags from the source booking system. An explicit opt-out received in a reply creates one automatically.
Reply Triage: Intent labeling of inbound replies — clinical, complaint, opt-out, cancellation, unknown — with a suggested reply that a human sends.; Reply triage never auto-responds. It labels the intent of each inbound message, attaches a suggested reply and a risk level, and routes clinical questions and complaints to a person. Opt-outs are recorded as contact exceptions immediately. Clients better served by a phone call enter a human call queue with a review record.
Support Session: A logged, purpose-bound, expiring window in which a TechStack operator can view a practice account read-only.; Platform staff cannot browse customer accounts at will. A support session records who opened it, the stated purpose, and its expiry; the operator works under a persistent read-only banner; and start and end events are written to the audit trail.

Want to see the math behind these terms running on your data?

15-minute walkthrough. We'll show you Book Value, Drift Score, and Voice Training computed against your actual book or pipeline.

Book a walkthrough Existing customer sign in

15 minutes · no sales pitch Works with your booking platform