Skip to content

Customers

Numbers, not quotes — and not until they're real.

This page will hold case studies with figures reconciled to Metrc and approved word-by-word by the store. Until founding stores have run long enough to produce them, here is what the product does on the demo tenant, exactly as it will do on yours.

DEMO TENANT · NORTHBANK CANNABIS · 3 STORES ON BLAZE + CSV · OVERNIGHT RUN RUN-0925-N

Drafted 5 · executed 0
  • 2m 14s

    to capture a 6-line invoice against the Metrc manifest; 1 line held on a tag mismatch, corrected-invoice request drafted

  • 21.7%

    of a weekday vape discount's redemptions fired outside its rule — $18,740 — with a POS rule fix drafted

  • −3.4 pts

    repeat rate for discounted first visits vs matched non-discounted (95% CI −6.3 to −0.5); proposal: hold the $120k Q4 increase

  • $3,044

    in vendor credit packs on Metrc-proven units, MI only; NY blocked by the inducement rule, CA routed to compliance

Every figure above is a fixture from the product specification, chosen to be internally consistent. It is not a customer result and is not presented as one.

The founding cohort

Stores listed here have agreed to be named. Results appear after a full pre/post benchmark from the ledger — never earlier, never estimated.

Join the cohort
  • Michigan

    Placeholder: [STORE NAME]

    Placeholder: [City] · Placeholder: [POS] · onboarding Placeholder: [MONTH]

    Benchmark in progress

  • California

    Placeholder: [STORE NAME]

    Placeholder: [City] · Placeholder: [POS] · onboarding Placeholder: [MONTH]

    Benchmark in progress

  • New York

    Placeholder: [STORE NAME]

    Placeholder: [City] · Placeholder: [POS] · onboarding Placeholder: [MONTH]

    Benchmark in progress

Placeholder: [N] more spots across MI, CA and NY — Apply →

How a case study gets onto this page

  1. 1

    Baseline from the ledger

    Four weeks pre-launch: tickets, AOV, repeat rate, recon match, Metrc lag.

  2. 2

    Post window, same metrics

    Twelve weeks live. Differences shown with a confidence interval or “not significant”.

  3. 3

    Store reads every word

    Written approval from the owner; anything they strike is gone.

  4. 4

    Published with its method

    Windows, sources and exclusions listed under every number.