It's the second Saturday in May. There are forty people in your hoop house, half of them holding a flat of something, and you are standing at a folding table trying to remember whether the 1-gallon coreopsis is priced the same as the 1-gallon salvia. Nothing in anyone's hands has a barcode on it. The tags are plastic stakes, half of them faded, and three of them fell out in the parking lot.
So you do what every small grower does: you eyeball the pot size, name a price, and take cash or wave someone toward a payment app. It works. It's also the reason you have no idea what actually sold.
Search for a fix and you'll find a dozen garden center POS systems. Read their pages closely and they all give you the same answer to "my plants don't have barcodes": print barcodes. Weatherproof plastic stakes, tree tags, laser-printed pot labels. Tag everything, then scan it.
That answer is not wrong. It's just written for a business roughly forty times your size.
Why nobody builds checkout for a nursery your size
The USDA's 2024 Census of Horticultural Specialties counted 23,060 horticulture operations in the US. Here's how they break down by annual sales:
Sixty-one percent of horticulture operations in this country sell under $100,000 a year. Together, that entire majority accounts for 3% of total horticultural sales. Meanwhile the largest 6% of operations — the ones selling $2.5M and up — bring in 75% of all sales.
If you sell POS systems, you chase that 6%. Of course you do. Those are the customers who can absorb a hardware package, a setup fee, a tag printer and an annual contract. The result is that the software written for "garden centers" assumes a receiving dock, a barcode label printer, and someone whose job is maintaining the item file.
If you're in the 61%, you're not a small version of that customer. You're a different business with a different problem: you need to ring up a plant in four seconds without having touched it beforehand.
What a no-barcode checkout actually looks like
Here's the flow when the item in someone's hands has no scannable code on it:
The important part isn't the camera. It's that you get three ways into the cart and can use whichever suits the item, instead of one way that demands every item be prepared in advance.
Be honest about what photo recognition can and can't do
This is where most vendor pages would keep selling. We'd rather you set it up correctly the first time, so here's the real boundary.
Photo recognition is genuinely good at telling apart things that look different. A Boston fern from a tomato start from a bag of mulch from a ceramic pot — no problem. It's identifying against your catalog, not the entire plant kingdom, which makes the job much easier than it sounds.
It is not reliable at cultivar-level distinctions in a black nursery pot. 'Green Velvet' boxwood and 'Winter Gem' boxwood are not going to be separated by a phone camera, and neither are twelve hosta varieties in June, or two hydrangeas that won't declare themselves until they bloom. Any product that tells you otherwise is selling you a disappointment in week two.
So use the right tool per item:
- Visually distinct stock — ferns, succulents, herbs, veg starts, hanging baskets, hardgoods: photo recognition earns its keep.
- Look-alike cultivars in identical pots — shrubs, hosta, daylilies, grasses: catalog tiles, or a barcode if it's high-value enough to justify a tag.
- Anything that arrived with a code on it: just scan it. Free accuracy.
- High-value specimen trees: tag them. A handful of tags is reasonable; four hundred is not.
The setup that actually saves you: price bands, not cultivar SKUs
Here's the thing most small growers get talked into and shouldn't: building an item file with one entry per cultivar.
You have 300 varieties. You do not have 300 prices. You almost certainly have about eight:
4" annual · 1-gal perennial · 2-gal shrub · 3-gal shrub · hanging basket · flat of 6 · bagged soil · specimen tree
Build your catalog as price bands first, and it takes an afternoon instead of a week. Every band is one tile on the screen. Checkout becomes "1-gal perennial ×3" and the math is right, the tax is right, and the card payment goes through — which is 90% of the value.
Then, once the season is running, promote the items you actually care about tracking into their own entries. The tomato starts you sell 400 of. The specimen Japanese maple. The things where knowing exactly what sold changes what you grow next year. Let the long tail sit in price bands where it belongs.
This is the opposite of how enterprise nursery POS onboarding works, and it's why those systems take a season to get live. You don't need a perfect item file. You need to stop losing sales at a folding table.
Where your sales actually happen
One more thing the USDA data makes clear. Of the direct-to-consumer channels — the ones where you handle the checkout rather than shipping to a wholesaler — selling on your own site dwarfs everything else:
On-site selling is nearly eight times the farmers-market channel and more than three times mail order. The checkout that matters most is the one standing in your own greenhouse — which is also the one most likely to be a folding table and a cash box.
If you do also load a truck for markets, the same catalog travels with you on the same device. And if your spring sale or fall festival charges admission, entry passes and printed wristbands can be switched on in settings too — that's covered in our u-pick and entry pass guide.
What about the greenhouse with no signal?
Hoop houses and back fields are famously bad for connectivity. Whether you can still take a card when the signal drops depends almost entirely on which card reader you chose — it's a hardware decision, not a software setting, and it's worth understanding before you buy. We wrote up the specifics in taking card payments when the internet goes down.
Getting started without a week of prep
If you want to try this before the next big Saturday, the shortest useful path is:
- Put your price bands in as catalog items — eight or so entries, one afternoon.
- Add real entries for your top 20 sellers, with a photo on each so tiles and photo-matching both work.
- Scan in your hardgoods — soil, fertilizer, tools — using the barcodes already on them.
- Stand a phone or tablet you already own on the potting bench and take a real sale.
That's enough to start. The catalog gets better as the season teaches you what you actually sell.
TallyTill runs on the phone or tablet you already have, and we publish our current pricing plainly on the pricing page rather than making you sit through a quote call. If you'd like to see it against your own plants, you can start a free trial and have a working till on the bench today.
Data in this article comes from the USDA NASS 2024 Census of Horticultural Specialties, published February 2026.