Tiller POS
Built to open the next store.
From the operator. I run a dispensary in Los Angeles, on Treez, five years of history. I'm building Tiller to open our second store on. Not a demo store. The real one, with my money on it.
Bring your last export. We'll show you what's in it.
01Who this is for
You're running Treez, Dutchie, Blaze or Flowhub. I was paying about a thousand a month for one of them. At some point in the last year you exported a report to a spreadsheet because the software wouldn't answer the question you actually had. You know your margin is thin. You don't know which product is thin and which one just looks thin.
If that's not you, this isn't your page.
02Who this is not for
We'll save you the call.
- Not for multi-state operators. No cross-state compliance, no org chart across licenses. One state, built for one state's rules.
- Not for anyone who needs a vendor to blame. No 24-hour desk, no account manager. There's an owner and a builder, and you'll talk to both.
- Not for anyone outside California. The tax engine and the Metrc rules are built to one state's law, on purpose.
- Not for anyone who wants to buy it and stop thinking about it. Tiller changes because someone standing at a register says what's wrong with it. If you won't say, we can't build you anything better than what you already have.
- Not for delivery or online checkout, today. Both are off by default. More on that below.
03Five Years at the Counter
When I switched systems, my inventory came across and my open tickets came across. Five years of knowing my regulars did not. Monday morning was a room full of strangers. Ask your vendor what they'll move for you, and hold the answer against what follows.
We rebuilt sixty-two months of the store's own ticket lines, July 2021 to August 2026, every line of every ticket, out of twenty-six Treez exports, and put the result on the budtender's screen. On August 28, 2026 the badge went live: REGULAR · away 10 months. The badge on the screen says how long, never how much, because the screen at the counter faces both ways.
The same rebuild found something worse: most of those customer records carried a placeholder email. The list is real, and today it's only reachable one way. At the counter. Which is where Tiller is built to win anyway.
(The store's own history, above. Yours will be your own.)
04Clean Before You Open
Here's what I hand you before you open the doors: the list of what's wrong with your Metrc data. What's negative, what's empty, what's flagged for recall, in your hands while you can still do something about it. Ask anyone else you're talking to for the same list, and ask when you'd get it.
We rebuild your inventory from Metrc, the record the state actually holds. On the last reload we ran, that was 63 negatives, 101 empties, 8 recalls, off a 172-item clean list.
Then we count. Blind, no expected number on the sheet, ordered by the zones on your own floor plan, so your crew walks the room instead of a spreadsheet.
Where a product name could mean forty different things on the shelf, we don't guess. We tell you it's a manager's call and say why. Refusing to guess is the feature.
Today, Metrc here means a manual, audit-first file, by design. We've applied for the vendor API key. We don't have it yet, so we don't claim a live connection.
05What Earns the Shelf
You already know your margin is thin. You don't know which product is thin and which one only looks thin.
We read your own tickets and tell you which product earns its place on the shelf, in dollars per unit, not percentages. Tiller's own reports carry your sales and your cost of goods by brand, by product, by category. The shelf verdict itself is work we do with you, from your own line-level history, not a button you press.
In our own store, that work found a category running the best percentage margin on the menu and one of the worst dollar returns per unit — and a different category earning three times as much per unit at a lower percentage. You only see that when the shelf is measured both ways.
Your best margin can be your worst shelf, and a percentage on a report will never show you that. Only your own tickets will.
06What we'll change the same day
On August 28, 2026, the operator asked at the counter for one new question at the register. It was in production that same afternoon.
We can do that today because there's one store and one builder behind this. We'll say plainly when we no longer can.
07What we don't do best, and who does
- Dutchie is the one to beat on ecommerce and payments.
- Treez is the one to beat on enterprise inventory and the number of things it plugs into.
- Flowhub's handheld scanner is well regarded.
We're not pretending otherwise. Tiller is not trying to be any of those, yet.
08Close
Bring your last export. Watch it load.
Fifteen minutes. You bring a Treez or Dutchie export, we show you what's actually in it, and you tell us if it's worth a second call.