A flower shop already has a printed interface. It is the care card in the sleeve, and it is handed over at the exact moment the customer stops needing the shop and starts needing instructions.
That card is also where most of the wasted design work in a shop goes, because it is asked to carry two facts with completely different clocks:
- "How do I keep these alive?" — the answer does not change for years. It is the same five steps in 2026 as in 2019.
- "What have you got today, and how much is a bunch?" — the answer was correct at 8am and wrong by lunch.
Print the second one and the shop ends up with a stack of cards for last week's peonies. Put the first one behind a code and the customer has to unlock their phone to be told to cut the stems — a medium problem solved with a screen it does not need. So the split is not "paper is old, screens are new"; it is sort the content by how often it changes, then give each clock the medium it deserves.
This is the model I ended up with for florists, and the part of it that took the longest was not the page. It was discovering which of the popular care instructions are actually true.
The five lines that are worth paper
Care cards fail by length. Every extra line reduces the chance the first line is read. Five lines fit on the back of a business card and each one maps to a documented cause of early collapse:
1. Cut 2 cm off each stem at an angle before it goes in the vase. 2. Clean vase, cool water two-thirds full, with the sachet. 3. Change the water and re-cut every two days. 4. No fruit bowl, no radiator, no direct sun. 5. Daffodils: own vase for the first day. Line 1 is the one everyone writes. Line 4 is the one that sounds like folklore and is not: ripening fruit gives off ethylene, which ages flowers and drops petals, so a vase next to a banana bowl really does die faster.
What the card should not carry is a number of days. Extension services put healthy cut flowers at ten to fourteen days under good conditions. A bunch that spent an afternoon warm in a van will not reach it, and the customer remembers the promise, not the conditions. In interface terms: don't render an optimistic estimate as a guarantee.
Where the folklore fails the test
This is the part worth writing up, because flower care has a documented gap between consumer advice and postharvest research, and the two most repeated tips sit on the wrong side of it.
| What everyone says | What the research finds | What to print instead |
|---|---|---|
| "Cut on a slant so they drink better." | The angle does not change vase life. Designers slant stems so they push into floral foam; the increased cut surface does not increase uptake. A blunt tool crushes the stem, and a crushed stem drinks badly. | "Cut 2 cm off with a sharp knife or scissors." |
| "Always cut stems under water." | Studies cited by the Society of American Florists found underwater cutting reduced vase life or at best matched a dry cut: microbes from cut stems collect in the reservoir and get drawn back into the next stems. | Nothing. Cut dry, with a clean tool. |
| "Warm water perks them up." | Water temperature alone does not extend vase life; warmth speeds opening, which is useful before an event, not for a vase at home. | "Cool water." Spring bulbs — tulips, daffodils, hyacinths — prefer cold. |
Consumer guides and the trade press genuinely disagree on the first two, and both are worth reading: extension services recommend a 45-degree cut because an angled end cannot sit flat and seal itself off, which is a good reason for an angle and a different reason from "it drinks better". The honest card states the action and drops the incorrect mechanism.
This matters more for a shop than for a blogger. A customer told to cut underwater, whose bunch fades in four days, does not conclude that the flowers were old. They conclude the shop gave bad advice — and that conclusion is about the shop, not the flowers.
The daffodil case, or why one line separates a florist from a card printer
Daffodil sap contains alkaloids and polysaccharides that damage other species. Unconditioned daffodils in a mixed bunch can shorten the vase life of the other flowers by up to half. The fix costs a day of foresight: stand freshly cut or re-cut daffodils in their own water for 12 to 24 hours before they join anything else. After that they can share a vase, and even be re-cut, without harming the rest.
So line 5 on the card — "Daffodils: own vase for the first day" — is five words of unusual, true, useful information. It also implies a production rule the shop has to follow: in spring, mixed bunches cannot be assembled the same morning the narcissus arrive.
Sorting the rest by clock speed
Once you accept the split, the rest falls out cleanly. Three tiers:
Yearly or never (printed): shop name, address, phone, the five care lines, the "no fruit bowl" line, opening hours, one line about standing orders.
Daily (page): which bunches exist right now, what is in each, this week's price, how many are left, what sold out since the morning.
Per-order (page or message): the same-day cutoff time and what it means — made today, or delivered today.
Two engineering consequences follow, and they are the only parts of this that are actually about QR codes:
- The printed code must point at an address, not at a file. A PDF or a photo of today's board gets a new URL every time it is replaced, so the printed code dies with the file. A page address survives every edit. This is the difference between "a QR code" and "a QR code that stays correct for three years".
- The printed card and the page must share the same strings. If the card says "change the water every two days" and the page says "every day", one of them is wrong and the customer will believe the paper one. Duplicate the five lines in both places, or generate both from one source.
The second point is where most multi-channel shops quietly diverge, and it costs nothing to prevent: keep the care lines as five strings in one file, render them into the card PDF and into the page.
Editing must be faster than answering the phone
A page behind a code has one failure mode that a website does not: it is trusted precisely where nobody can see it going stale.
If marking two bunches sold out takes longer than answering the phone, it will not get done, and a page that is three days stale is worse than no page — the code on the sleeve then actively misleads. The test before ordering a thousand cards is not "does the page look good", it is: can the person at the counter mark something sold out in under a minute while a customer waits?
That constraint also drives a design decision: keep sold-out items on the list with a mark instead of deleting them. The customer who drove over for peonies needs to see that the peonies are gone, not wonder whether they misread the board.
Printing a code onto a surface that gets crushed
A bouquet sleeve is the worst surface in the shop for a code: crinkled, curved, often transparent, handled with wet hands. Four rules that survived contact with reality:
- Keep it off the fold. A code crossing a crease loses modules exactly where the paper bends. Four modules of clear space around the code is the usual quiet-zone requirement; a seam inside it will eat it.
- Print the code on a card, not on the sleeve. A small card tucked into the sleeve stays flat and outlives the wrap — which is precisely when the care questions begin.
- Size by reading distance. One unit of code per ten units of distance is the industry rule of thumb: 20 cm away needs about 2 cm of code, 50 cm needs 5 cm. It is a floor rather than a promise, and it is not part of the QR standard, so behind reflective glass plan for one per eight instead of one per ten.
- Print the address in text under the code. If the ink smears or a corner is torn off, a page someone can type is still a page.
In our own kit the counter card is 93.0 × 127.5 mm with a 48 mm code (reads at about half a metre) and the A4 poster carries a 126 mm code (about 1.3 m). Those numbers come from measuring the generated PDFs, not from estimating them — and both assume matte stock, because a glossy sleeve behind a shop window will reflect a streetlight into the scanner.
What this looks like as a page
Eight blocks, in this order, and nothing else:
- Today's bunches — name, contents, price, how many left
- Same-day cutoff — one time, plus what "same day" means
- The five care lines — identical to the card
- Vase life by flower type — with the honest "a mixed bunch lasts as long as its shortest-lived stem"
- Standing orders — with the pause rule in one line
- Events, funerals, weddings — separate line, separate form
- Where and when — address, hours, parking
- One button — call, message, or order. One, not four.
Most disappointing flower orders are a timing problem, not a quality problem: someone asks for delivery "today" at 4pm, the shop can make it but cannot deliver it, and the answer arrives too late to fix anything. The cutoff, printed where the customer sees it before asking, is the most valuable line on the page and costs one sentence.
I build these pages and print the QR pieces that go with them: a free one-page shop site, and a one-off printed kit — the florist guide, the $49 kit for a shop with no tables (most florists: a page plus eight files), and the three-year cost calculator for comparing a subscription against a one-off. If you only want the printing geometry, the QR size guide in millimetres has the floor, the quiet zone and the 1:10 rule laid out.
Care facts here come from extension services (NC State, UMass, University of Illinois), the RHS, UC Davis postharvest notes and the Society of American Florists' trade magazine — dates and titles are in the sources block of the guide page. Vase life depends on variety and cold chain, so none of it is a promise about a particular bunch.
